1. Investment Summary

We believe ETH represents an outlier opportunity over the next two to three years. The market still values Ethereum at a bear-market trough multiple of the value the network secures, while the fundamentals that drive that value compound, and our probability-weighted 2028 scenarios imply approximately $9,700 per ETH, more than four times the current price.

ETH closed July 2026 at approximately $1,863, about one-third below its December 2025 level, and has since repriced sharply: as of August 20 it trades near $2,280, following the Treasury’s expanded bond buybacks, the SEC’s proposed Regulation Crypto Assets, and a $2.9 billion short liquidation. We read this move as the beginning of a larger repricing rather than the end of one, because the gap between Ethereum’s fundamentals and its valuation remains wide.

While the price fell through the first half of 2026, the fundamentals compounded. Tokenized real-world assets (”RWAs”) on the network roughly tripled over the past year, the United States established a federal stablecoin framework, regulators cleared staking inside ETFs, DTCC moved tokenized securities into live trading, and Visa and Mastercard joined a payment standard built for AI agents.

The thesis rests on two reinforcing adoption curves. First, traditional finance is moving onchain: tokenized RWAs grew from ~$12 billion to $33.5 billion in twelve months, and Ethereum settles roughly half on its base layer, over two-thirds including its L2s, with BlackRock, DTCC, Robinhood, and Cantor Fitzgerald now in production. Second, AI agents require programmable financial infrastructure: an autonomous agent cannot open a bank account, but it can hold a wallet, verify counterparty code, and settle in seconds, which makes public blockchains the natural venue for machine-driven commerce.

We value ETH through the security ratio: ETH market capitalization relative to the value secured on Ethereum. Even after the August rally, the ratio stands near 1.2x, a level associated with bear-market troughs across the past six years, implying the market still assigns little value to future growth. Our 2028 scenarios rely on stablecoin, tokenization, and DeFi growth only; AI-agent adoption supports a separate fee case.

2. Why Now

The first half of 2026 delivered a full deleveraging: total crypto market capitalization fell ~47% from its October 2025 peak, ETH traded below $1,800, and US spot ETH ETFs saw ~$4.4 billion of outflows. Positioning reached its lightest level since the ETFs launched, while every adoption metric moved the other way. Ethereum’s base-layer RWA value rose ~300% year over year to over $17 billion, stablecoins crossed $300 billion with roughly half settling on Ethereum, and the network processed a record 200 million L1 transactions in the first quarter.

The repricing has now begun. ETF flows turned positive in mid-July, the ETH/BTC ratio bottomed, and the August 19 rally added ~20% in a day on record buying volume. What has not repriced is the valuation multiple: at ~1.2x value secured, ETH still trades at trough levels while regulation, institutional deployment, and the supply structure have all improved. Windows between the fundamentals turning and the multiple normalizing are where the asymmetry lives.

3. Two Reinforcing Adoption Curves

Traditional finance is moving onchain. Institutions are no longer buying crypto; they are issuing their own products on public rails, because tokenized assets settle in seconds, trade continuously, serve as programmable collateral, and embed compliance in software. Tokenized Treasuries and money-market funds reached ~$13 - 15 billion, led by BlackRock’s BUIDL. Robinhood offers 2,000+ tokenized equities across 120 countries on its own Ethereum-ecosystem chain. DTCC trades tokenized securities live, and Cantor runs blockchain-based IPOs with Securitize. Stablecoins, at ~$310 billion and settling volumes comparable to card networks, show what the mature version of this curve looks like; Treasuries, funds, and equities are following the same path. Settlement concentrates where liquidity and collateral already sit: Ethereum holds ~48% of tokenized RWAs on L1, over 66% including L2s, and ~half of all stablecoins. Our base case assumes sharp deceleration with ~$500 billion of RWAs by 2028 against street forecasts of $2–16 trillion by 2030.

AI agents require native financial infrastructure. An agent in traditional finance cannot satisfy KYC, hold assets in its own name, or verify what an intermediary will do; onchain it creates a wallet in milliseconds, inspects counterparty code, and composes trading, lending, and settlement into one atomic workflow on markets that never close. Tokenization puts assets where agents can reach them; agents create transaction demand tokenization alone would not. The early evidence is directional: Coinbase’s x402 protocol processed 165 million+ transactions within months, Visa, Mastercard, and Ripple joined the standard in July, and Google’s AP2 counts 60+ partners, with most activity settling in USDC on Base, an Ethereum L2. Our valuation scenarios do not require agent adoption; agents underwrite the separate fee-recapture case as Glamsterdam expands L1 capacity.

4. Why Ethereum

Neutrality and security are underwriting facts, not ideology. Ethereum has no corporate owner, five-plus independent clients, a permissionless validator set in the hundreds of thousands, and continuous uptime since 2015; every other major L1 reviewed in the Ethereum Foundation’s institutional risk assessment has halted at least once, and one competitor’s corporate backer controls ~42% of token supply. Roughly $76 billion of staked ETH secures the network. Finalizing a fraudulent transaction would cost an attacker tens of billions and destroy the stake. No competing smart-contract network approaches this, and as tokenized value moves into the hundreds of billions, the security behind settlement becomes an explicit institutional constraint.

The operating record is unmatched. Ethereum has rebuilt its own engine mid-flight including replacing its consensus mechanism without a minute of downtime, and now ships major upgrades annually (Pectra, Fusaka delivered on schedule; Glamsterdam in flight). With 11,000+ active EVM developers, it is the only ecosystem formally working on post-quantum cryptography and institutional privacy at protocol level. For a ten-year infrastructure decision, the question is which network remains secure, neutral, and adaptable as requirements change; Ethereum has the strongest record against that standard.

The institutional machine has been parallelized. The Foundation cut its budget ~40%, refocused on protocol stewardship, and stakes rather than selling its treasury. Around it, specialized and independently funded entities now own the go-to-market: Etherealize (institutional tokenization and policy, $40M from Electric Capital and Paradigm), Ethereum Institutional (the entry point for banks and sovereigns), EthLabs (independent R&D), and EthSystems (bank-grade privacy). The Foundation’s own Ethereum for Governments and Institutions now markets the network to states directly.

The roadmap attacks the one real weakness. L1 fees collapsed as activity moved to L2s. Glamsterdam, targeted for late 2026, enables parallel execution and a path toward ~10,000 TPS on L1, while Fusaka’s blob-fee floor already prices L2 data availability into ETH burn. Glamsterdam creates capacity; agents are the credible demand that fills it.

5. Market Structure: Productive Demand, Reduced Float

Digital asset treasury companies hold ~7.9 million ETH, 6.5% of supply. BitMine alone holds 5.79 million after 53 consecutive weekly purchases, and stakes 4.9 million of it at ~2.8%, generating over $300 million of annualized income at current prices. This is internal cash flow that a non-yielding treasury asset cannot produce, which is why these vehicles have held through the drawdown without becoming forced sellers. They also fund the ecosystem’s institutional entities, aligning 6%+ of the ownership base with adoption. DAT equities trade at or below NAV; the observed response has been accretive buybacks rather than asset sales, and low leverage limits the unwind risk.

ETFs hold ~$12 billion and, since the March 2026 staking classification, pay yield, BlackRock’s ETHB nets ~2%. With ~30% of supply staked, 6.5% in treasuries, ~5% in ETFs, and net issuance of 0.2 - 0.3% (~$0.6 - 0.8 billion at current prices), identifiable structural demand of $10–20 billion annually in 2027 - 28 meets a thin float. Flows do not set the destination, but they set the speed and the past week demonstrated the mechanism.

6. Economics and Valuation

Ethereum sells programmable trust: the ability for any two parties, human or machine, to transact with finality without an intermediary. Fee-based valuations broke when scaling succeeded and fees collapsed; the relationship that has held through both fee regimes is between ETH’s market capitalization and the value the network secures. A settlement network is credible only if corrupting it costs more than the theft yields, and Ethereum’s security is denominated in staked ETH, so the cost of attack scales with ETH’s market capitalization.

We define the security ratio as ETH market capitalization divided by value secured (stablecoins ~$154B + tokenized RWAs ~$17B + DeFi TVL ~$56B ≈ $227B). At today’s ~$276 billion market capitalization the ratio is ~1.2x.

A daily proxy reconstructed from January 2021 (ETH market capitalization over Ethereum stablecoins plus DeFi TVL) ranged from ~1.0x to 3.7x with a median of 2.2x. Bear-market troughs clustered at 1.0 - 1.3x; bull markets reached 2.4 - 3.7x. Today's ~1.2x sits in the trough band. The scenario ratios follow from this record: the bear case's 0.8x is below any sustained level, the base case's 1.3x is normalization to the top of the trough band, still well below the median, and the bull case's 1.5x sits well below the 2.2x median and far below prior bull-market peaks.

The operating forecast moves value secured from ~$227 billion to ~$400 billion in 2027 and ~$800 billion in the 2028 base case, driven by stablecoins compounding at ~40% annually at flat Ethereum share, RWA growth decelerating to ~$500 billion globally, and DeFi TVL reaching ~$150 billion. Post-Glamsterdam, we expect paid L1 throughput of 2,000 - 5,000 transactions per second by 2028, predominantly agent-driven, producing $0.9 - 3.2 billion of annual fee revenue as a cross-check. Supportive, but not what carries the valuation.

The bear case deserves attention for what it requires: stablecoins growing only 45% in two and a half years despite federal law, RWAs growing 4.5x after 2.8x in the past year alone, Ethereum losing share everywhere, and the multiple de-rating below its six-year floor and the result is approximately today's price. The base case requires no heroics and lands at $8,600, well below Standard Chartered's $25,000 target for 2028. The bull case adds moderate multiple expansion: at the same $1.55 trillion secured, the base-case 1.3x ratio alone would imply ~$16,700, and the re-rating to 1.5x carries the case to ~$19,200 while remaining well below the historical median. The probability-weighted value is ~$9,700, or ~4.3x.

Sensitivity separates the two sources of upside, adoption (rows) and re-rating (columns):

We report value secured, the security ratio, Ethereum’s RWA and stablecoin share, L1 and blob revenue, ETF staking assets, DAT flows, and agentic transaction activity in every investor letter.

7. The Competitive Landscape: A Frontier and Not a Race

The most performant blockchain possible runs on a single computer; the most trustworthy runs on every computer on earth. Neither extreme clears a market, and every network chooses a position between them. Demand pools where use-case requirements match that position, and four durable pockets have formed: Hyperliquid at the performant end, purpose-built for perpetuals; Solana next, with coordinated infrastructure serving high-velocity consumer activity; Ethereum and its L2s in the decentralized range, hosting the assets for which security and neutrality bind; and Bitcoin at the far end, serving exactly one use case, hard money. None is mispositioned but transaction counts concentrate at the performant end, while balance sheets concentrate at the trust end, and our valuation framework measures balance sheets.

Institutional demand has already revealed where it lands. The permissioned consortium chains of 2016 - 2019 offered maximum performance and control and failed commercially, because settlement without open liquidity and neutral governance is a database with extra steps; the banks that funded them now deploy on public Ethereum, at several times the share of any alternative, despite cheaper blockspace being freely available. Settlement also compounds: each institution that arrives makes every incumbent’s assets more useful as collateral and liquidity, raising switching costs.

The positions are not equally mobile. Performance is an engineering problem and decentralization is a history problem. Ethereum can move toward the performant end and Glamsterdam does exactly that at constant decentralization. While no performant chain can retrofit a decade of uptime, a $76 billion security bond, or credible neutrality within an investment horizon. Through its L2s, Ethereum is also the only ecosystem selling a range of the frontier rather than a single point. Competitors sell one tradeoff; Ethereum sells the curve.

8. Risks and Disconfirming Evidence

The thesis is designed to be falsifiable. These thresholds trigger formal re-underwriting rather than being treated as volatility:

The most important risk is value-capture failure: activity migrating to L2s that pay the base layer little, leaving ETH widely used but not held. Our counterview is that ETH remains the network’s security, collateral, staking asset, and gas reserve; a growing pool of tokenized value cannot rely indefinitely on an undercapitalized security layer, and Fusaka has already shown the protocol will charge L2s. Regulatory delay would flatten the adoption curve without bending it back. The GENIUS Act takes full effect by January 2027 regardless, and DTCC, BlackRock, and Robinhood built under existing authority. We also monitor DAT reflexivity, competitive share loss, and staking concentration, each observable and reported.

9. The Dialectic Position

Dialectic has operated natively onchain across multiple cycles. The strategy captures both the re-rating and the yield the convergence creates: tokenized assets need liquidity, staking and restaking carry structural spreads, and DeFi credit markets misprice during regime shifts. The fund holds core ETH exposure and deploys onchain yield strategies around it. In the base case, yield compounds on appreciation; in a flat market, it pays for patience. Dialectic’s relationships across protocols, founders, and the core Ethereum organizations provide direct visibility into tokenization mandates and emerging risks; the monitoring framework in Section 6 is how the strategy is already managed.

The market values ETH at ~1.2x the value Ethereum already secures, a trough multiple, while tokenized assets, stablecoins, institutional infrastructure, and staking products compound. The bear case resolves near today’s price, the base case supports ~$8,600, and the probability-weighted value is ~$9,700. August showed how quickly the gap can start to close. We are positioned for the rest of it.

Disclaimer

This memorandum is for informational purposes only. It does not constitute investment advice, a recommendation, or an offer to buy or sell any security or digital asset, and it is not a solicitation to invest in any Dialectic fund or Machine. It reflects Dialectic’s own research and views as of August 2026 and does not account for any reader’s individual financial situation or objectives.

Scenario values, price levels, and probability weightings in this memorandum are estimates and internal models, not forecasts, targets, or guarantees; actual outcomes may differ materially, and Dialectic does not guarantee any yield or rate of return. Digital assets are volatile and can lose some or all of their value, and past performance is not indicative of future results.

Recipients should conduct their own due diligence and consult independent financial, legal, and tax advisors before making any investment decision.

This report covers the performance and positioning over the past few weeks of DUSD, DETH, and DBIT. Inside, you will find a clear view of performance drivers, allocation shifts, and risk management decisions across each Machine.

DUSD

The DUSD Machine generated a 14-day average yield of 8.18% on Meccanico’s deployed capital over the past two weeks. The uplift in returns was driven primarily by 3Jane Ecosystem Vault on Morpho, USDG on Aave with an additional contribution from the monthly NAV update of the Pareto FalconX position.

On the allocation side, the RLUSD Sentora vault was closed during the period, with the Machine rotating into two new positions: a PRIME/PYUSD loop and the Dialectic AUSD RWA vault on Morpho, which generates yield by lending AUSD against a curated set of tokenised credit and real-world-asset collateral focused on short-duration fixed-income assets.

In summary, the DUSD Machine continues to focus on high risk-adjusted RWA opportunities with tactical allocations to looping while maintaining high liquidity.

DUSD Summary:

DUSD Performance:

DUSD Asset Allocation Breakdown:

DUSD Protocol Allocation Breakdown:

DUSD Chain Allocation Breakdown:

DUSD Exposure:

DETH

The DETH Machine generated an average 3.62% APR on Meccanico’s deployed capital over the reporting period, with performance improving through the fortnight. Week 30 came in at 2.99% APR and week 31 rose to 4.23% as the levered staking carry benefited from a firmer yield-to-borrow spread. The onchain price of the DETH moved from 1.016078 to 1.017489 ETH, and total value deployed ended the period at 2,592 ETH (around $4.8M).

Positioning remains anchored by the levered LST loop, which runs roughly 12,560 ETH of gross collateral against 11,320 ETH of WETH debt at about 10x, for a net exposure of 1,246 ETH or 48% of the Machine’s deployed capital. A further unlevered yield sleeve holds 19% (around 497 ETH), and 12% sits in the wstETH liquidity sleeve of 252.8 wstETH, which the DETH Machine maintains to meet redemptions and rebalance without touching core positions.

The most significant development of the period was the scaling of Meccanico’s allocation to the DQaeETH arbitrage machine, from roughly 248 ETH to 556 ETH, now around 21% of the Machine’s deployed capital. The strategy was still ramping up through the reporting period as Meccanico tuned execution and added more swap integrations to source the best price quotes across venues.

The results are starting to show, with the trailing 7-day yield now annualizing at close to 5% APR and continuing to climb as cycles compound. DQaeETH runs an ETH-native arbitrage on the ether.fi ecosystem, accumulating weETH when it trades at a discount to NAV in secondary markets and redeeming it at par through the withdrawal queue. Since the return comes from the discount captured rather than from staking rates, borrow costs, or ETH price direction, its P&L is structurally uncorrelated to the levered carry that drives the rest of the Machine’s positions, while remaining fully ETH-denominated at every step. It gives the DETH Machine a second, independent return source. When staking spreads compress and the loop’s carry softens, LRT discounts typically widen, improving arbitrage economics exactly when the Machine needs it most.

Net redemptions of roughly 73 ETH (2.75% of supply) were serviced entirely from the liquidity sleeve. With the week 31 run-rate back above 4%, the yield compression flagged in the weeks 26–27 update has partially reversed, and Dialectic Meccanico continues to deploy excess liquidity into low-risk, short-duration opportunities to minimize cash drag.

DETH Summary:

DETH Performance:

DETH Asset Allocation Breakdown:

DETH Protocol Allocation Breakdown:

DETH Chain Allocation Breakdown:

DETH Exposure:

DBIT

DBIT Machine generated a 2.67% APR on Dialectic Meccanico's deployed capital in the past 14 days; returns were driven primarily by the Across LP allocation. At the same time, the Machine's positions have since diversified toward a conservative carry strategy leveraging the Senotra PYUSD vault and the WBTC/USDC Morpho Market (both comprising 56.7% of the DBIT Machine's positions). The remaining portion of the Machine's positions are allocated to the Across LP.

DBIT Summary:

DBIT Performance:

DBIT Asset Allocation Breakdown:

DBIT Protocol Allocation Breakdown:

DBIT Chain Allocation Breakdown:

DBIT Exposure:

Thanks for reading.

To learn more about Dialectic.

For the latest news, follow us on X.

DUSD, DETH, and DBIT Machines are being operated on Makina - the DeFi execution engine.

Have questions or comments? Head over to Discord and open a ticket or you can also find us in Makina’s Telegram.

Disclaimer

Machine Tokens (DUSD, DETH, DBIT) are products sold by Dialectic Meccanico Ltd. (BVI) as principal seller for its own account and are issued by the relevant Machine smart contract on the Makina Protocol; they are not deposits, fund interests or participations in a collective investment scheme, and Meccanico operates each Machine with its own capital rather than managing assets for Machine Token holders.

The Machine Tokens are sold “AS IS, WITH ALL FAULTS“, their value may fall to zero, and there is no redemption right, as any Buyback is made at Meccanico’s sole discretion and may not occur; no value, return or performance is guaranteed, and past strategy performance is not indicative of future strategy performance.

This material is informational only, is not part of the applicable Makina terms and/or Dialectic Meccanico Ltd. (BVI) Operator Terms, and is not investment, tax or legal advice or any recommendation, solicitation, endorsement or offer to acquire, hold or dispose of any Machine Token, or acquire, hold or dispose of any investment product, vehicle, service or instrument.

Past strategy performance is not indicative of future strategy performance.

This report covers the performance and positioning over the past few weeks of DUSD, DETH, and DBIT. Inside, you will find a clear view of performance drivers, allocation shifts, and risk management decisions across each Machine.

DUSD

DUSD Summary:

Over the past two weeks, the DUSD Machine, which executes Meccanico’s proprietary USD strategy on the Makina Protocol, generated an annualized yield of approximately 4.92% on Meccanico’s deployed capital. The slightly lower returns are explained by the execution costs from closing the remaining concentrated loops.

Other major changes included the closure of the 3Jane USD3 senior tranche and the Sentora PYUSD vault and rotation into a new position: USDG on Aave V4. In addition to that, the DUSD Machine increased its exposure to the Pareto FalconX position and allocated to the 3Jane Morpho ecosystem position which is heavily incentivized and offers better risk-adjusted returns compared to 3Jane’s senior tranche.

In parallel, integration work continued around Monad, with updates rolled out to support the Machine’s expanding deployment there.

Taken together, the period’s changes continue the deliberate rotation of the Machine’s positions towards real-world-asset (RWA) yield sources with its private credit RWA exposures while retaining ample liquidity throughout.

DUSD Performance:

DUSD Asset Allocation Breakdown:

DUSD Protocol Allocation Breakdown:

DUSD Chain Allocation Breakdown:

DUSD Exposure:

DETH

DETH Summary:

The DETH Machine’s proprietary ETH strategy on the Makina Protocol generated the following results over the last period, generating an annualized yield of approximately 4.10% on Meccanico’s deployed capital. This represents a notable increase from the 3.0% recorded in the previous period. The on-chain price of the Machine Token set by the relevant smart contract moved from 1.014724 to 1.016321 WETH/DETH over the period, alongside a TVL of approximately 2,662 WETH. This performance underscores the strategy, which prioritizes captured spreads over broader market beta.

Operational efficiency within the engine continues to improve. The primary leveraged wstETH/ETH loop successfully capitalized on easing ETH borrowing costs to capture wider spreads. Furthermore, diversification efforts are progressing according to plan; a carry sleeve was initiated during this window, and the DQaeETH arbitrage allocation is now fully deployed. These additions reduce reliance on individual positions, developing the Machine into a more resilient yield engine.

The Machine’s 30-day trailing yield is approximately 3.49% on Meccanico’s deployed capital, with a current run rate in the mid-3% range. Meccanico anticipates that some of the more recently added yield drivers have yet to contribute fully, subject to operational and market conditions and ongoing risk assessment; the carry sleeve is still in its early ramping phase, and the DQaeETH ending has been optimized further to capture better spreads.

DETH Performance:

DETH Asset Allocation Breakdown:

DETH Protocol Allocation Breakdown:

DETH Chain Allocation Breakdown:

DETH Exposure:

DBIT

DBIT Summary:

No major changes have occurred in the DBIT Machine, which executes Meccanico’s proprietary wBTC strategy on the Makina Protocol (annualized yield of approximately 1.28% on Meccanico’s deployed capital in the first half of July), with the Machine’s positions still concentrated in the Across LP pool, which currently offers a positive risk-adjusted wBTC yield profile.

DBIT Performance:

DBIT Asset Allocation Breakdown:

DBIT Protocol Allocation Breakdown:

DBIT Chain Allocation Breakdown:

DBIT Exposure:

Thanks for reading.

To learn more about Dialectic.

For the latest news, follow us on X.

DUSD, DETH, and DBIT Machines are being operated on Makina - the DeFi execution engine.

Have questions or comments? Head over to Discord and open a ticket or you can also find us in Makina’s Telegram.

Machine Tokens (DUSD, DETH, DBIT) are products sold by Dialectic Meccanico Ltd. (BVI) as principal seller for its own account and are issued by the relevant Machine smart contract on the Makina Protocol; they are not deposits, fund interests or participations in a collective investment scheme, and Meccanico operates each Machine with its own capital rather than managing assets for Machine Token holders.

The Machine Tokens are sold “AS IS, WITH ALL FAULTS”, their value may fall to zero, and there is no redemption right, as any Buyback is made at Meccanico’s sole discretion and may not occur; no value, return or performance is guaranteed, and past strategy performance is not indicative of future strategy performance.

This material is informational only, is not part of the applicable Makina terms and/or Dialectic Meccanico Ltd. (BVI) Operator Terms, and is not investment, tax or legal advice or any recommendation, solicitation, endorsement or offer to acquire, hold or dispose of any Machine Token, or acquire, hold or dispose of any investment product, vehicle, service or instrument.

In brief:

• Any DeFi yield comes with two options: deposit and hold all of its risk, or pass and earn none of it.

• Tranching adds a third. It splits the strategy into a protected Senior position and a first-loss Junior position, each priced and held on its own.

• Allocators can now stop choosing whole positions and start holding the exact slice of risk they want.

• The same strategy then reaches more capital and risk carries a clear price.

• Our view is tranching is turning from a product into infrastructure. Below we set out how it works and where Dialectic Meccanico already uses tranching to curate a strategy specific vault.

Splitting out risk into options

A standard DeFi vault offers one risk level. Conservative capital passes, because it eats the full loss if the strategy draws down. Aggressive capital passes too, because willingness to absorb that loss earns them nothing extra.

Tranching splits a single yield into two products, each for a different investor:

• Senior, the protected side: drawdown protection down to a set floor, at a lower rate. This is for capital that has to protect principal.

• Junior, the paid-to-take-risk side: first in line for any loss, and paid a premium out of Senior’s yield for this.

The asset and the strategy underneath do not change, only the risk is repackaged. On Avalanche (at the time of writing) Avant’s savUSD earns about 7.95%, split into a Senior near 7.57% and a Junior near 9.23%, with Senior protected down to a 20% floor. Junior’s extra yield is Senior’s payment for the protection. The same split now runs across many Royco Dawn markets.

Risk gets a price

In a blended vault, the price of risk is buried inside a single APY. Tranching splits it open through two numbers:

• The Junior premium rises and falls with demand for first-loss capital.

• The protection floor is sized to each strategy, higher for riskier ones, lower for steadier ones. Today it runs roughly 3% to 20% across live Royco Dawn markets.

Together, these numbers tell a curator/allocator what a position’s risk is worth before committing to it.

This clarity is what gives a strategy manager options depending on the mandate:

• Capital that must protect principal takes Senior: covered exposure with a defined buffer, underwritten by Junior.

• Capital paid to underwrite risk takes Junior: a standalone first-loss product, and with higher yields.

Tranching makes first-loss a product an allocator can actually price and choose.

A Senior vault curated by Dialectic Meccanico

Senior tranching is not a single-market idea. The Senior Royco USDC Vault (srRoyUSDC) already spreads Senior exposure across many Royco Dawn markets.

Because the book holds Senior tranches from several different issuers, a problem at any one of them touches only a part, not the whole vault.

Senior is a new kind of collateral

Because Senior carries drawdown protection down to its floor, it is considered less risky than the raw strategy behind it, so a lending market can lend against it at a higher LTV than against that raw asset. Senior becomes its own category of collateral. It supports more borrowing and keeps earning while it is posted as collateral.

For curators like Dialectic Meccanico, that opens a strategy the raw asset never could. Dialectic Meccanico can loop Senior as safer collateral, borrow against it, and redeploy, building a mezzanine layer that is leveraged but sits behind drawdown protection the whole way down. The leverage lifts the return and the coverage limits the downside on each underlying position.

Losses are handled with discipline

For institutions, the way losses are handled decides whether a position can be held at all. Two things matter:

• A short drop does not trigger a loss. When a strategy dips, the market enters an Observation Period. Junior only takes a loss if the drop lasts past a set time or goes beyond a set size. A brief dip that recovers does not count.

• Pricing follows the real value of the assets and not a temporary market price. So a short-lived dislocation doesn’t force a markdown and the protection floor is enforced by the contract.

From product to infrastructure

Tranching does not care what the strategy is. Any yield-bearing asset with a priceable value can be split, and the live Royco Dawn markets already include credit-backed stablecoins, delta-neutral strategies, and real-world-asset yield.

Royco standardizes the contracts, the protection enforcement, and the premium routing across all of them, so for an issuer, launching a tranched market is a distribution decision and not a rebuild.

The same is true on the curation side. Allocating the Senior Royco USDC Vault across tranched markets is active work, and Dialectic Meccanico does it through Makina’s onchain execution infrastructure, with every permitted action predefined onchain. Royco standardizes the product layer and Makina’s vault infrastructure gives the curator the operational layer to effectively manage the vault.

Why this matters

For decades, only traditional credit markets could slice risk into tranches. Now DeFi can too, but with the pricing in the open and the settlement automatic.

For Dialectic Meccanico as the curator, the payoff is direct. Strategies that were off-limits are now considered for inclusion in a strategy, with their risk priced in plain sight.

This is the whole idea behind risk as a product. Tranching turned risk from something a depositor holds into something an allocator chooses. If the Senior side is the side you’d choose, the srRoyUSDC vault is live.

All vault activity carries risks including smart-contract risk, oracle risk, market risk, liquidity risk, and protocol-governance risk. Past strategy performance is not indicative of future strategy performance. Investments involve risk, including potential loss of principal.

Nothing in this post should be considered investment, tax or legal advice or the recommendation to sell, or the offer of a solicitation to buy or invest in any investment product, vehicle, service or instrument.

This report covers the performance and positioning over the past few weeks of DUSD, DETH, and DBIT. Inside, you will find a clear view of performance drivers, allocation shifts, and risk management decisions across each Machine.

DUSD

Over the past few weeks, the DUSD Machine generated a 6.32% APR, with meaningful contributions from Sentora’s Morpho vaults and tokenized receivables through 3Jane. The period’s focus remained shifting exposures away from concentrated looping and towards real-world-asset (RWA) sources. The Convex USDT position was wound down, loop exposure was reduced further, and the allocation to 3Jane’s senior tranche and PYUSD, was increased.

Taken together, the period’s changes continue the deliberate rotation of the Machine’s positions. With loop exposure now largely unwound, the Machine enters the next period positioned around a more diversified, less rate-sensitive yield base while retaining ample liquidity throughout.

DUSD Summary:

DUSD Performance:

DUSD Asset Allocation Breakdown:

DUSD Protocol Allocation Breakdown:

DUSD Chain Allocation Breakdown:

DUSD Exposure:

DETH

The DETH Machine generated an average 3.02% APR over the reporting period. Performance was modestly lower than in the previous two weeks, primarily due to a softer weETH print, which resulted in a slight compression of the strategy’s overall yield.

During the period, Dialectic Meccanico also completed an allocation from DETH into DQaeETH, EtherFi’s ETH arbitrage strategy. The capital is now fully deployed and is currently awaiting redemption fulfillment, after which the first arbitrage cycle is anticipated to begin, subject to market and operational conditions.

In parallel, Dialectic Meccanico has begun building out the carry trading sleeve, focusing on opportunities that offer positive risk-adjusted returns while complementing the Machine’s existing sources of yield. Initial allocations include strategies such as the Sentora PYUSD/RLUSD Morpho Vault and the recent Aave Monad campaign, with additional opportunities expected to be added as they arise.

Beyond the incremental yield, these allocations improve the Machine’s overall capital efficiency. Maintaining a liquidity sleeve in wstETH is a deliberate design choice, ensuring sufficient flexibility to meet redemptions and rebalance positions as market conditions evolve. By opportunistically deploying excess liquidity into low-risk, short-duration opportunities with attractive risk-adjusted returns, we can meaningfully reduce cash drag and enhance portfolio carry without compromising liquidity or execution flexibility.

DETH Summary:

DETH Performance:

DETH Asset Allocation Breakdown:

DETH Protocol Allocation Breakdown:

DETH Chain Allocation Breakdown:

DETH Exposure:

DBIT

DBIT generated 0.78% APR in the reported period. Noting the persistent borrow rate volatility on Morpho and AAVE, the DBIT Machine has concentrated on Across LP provisioning. This decision has been made to address the NAV volatility of the previous weeks. That said, the Machine remains on the lookout for stablecoin carry opportunities as the borrow rates stabilize.

DBIT Summary:

DBIT Performance:

DBIT Asset Allocation Breakdown:

DBIT Protocol Allocation Breakdown:

DBIT Chain Allocation Breakdown:

DBIT Exposure:

Thanks for reading.

Click to learn more about Dialectic.

For the latest news, follow us on X.

DUSD, DETH, and DBIT Machines are being operated on Makina - the DeFi execution engine.

Have questions or comments? Head over to Discord and open a ticket or you can also find us in Makina’s Telegram.

Risk and Disclaimer:

Each Machine Token (DETH, DBIT, DUSD) is a product sold by Dialectic Meccanico Ltd. (BVI) as principal seller, acting for its own account. Each Machine Token is issued at the protocol level via the Makina Caliber smart contract on the Makina Protocol when the purchase price is paid. Machine Tokens are sold on an “AS IS, WITH ALL FAULTS” basis; users bear full risk, including total loss of the amount paid. Dialectic Meccanico Ltd. (BVI) operates the underlying Machines using its own capital. Machine Token holders participate at the protocol level; Meccanico does not manage assets for, or act on behalf of, Machine Token holders.

All Machine activity carries risks including smart-contract risk, oracle risk, market risk, liquidity risk, and protocol-governance risk. Past strategy performance is not indicative of future strategy performance. Investments involve risk, including potential loss of principal.

Nothing in this update should be considered investment, tax or legal advice or the recommendation to sell, or the offer of a solicitation to buy or invest in any investment product, vehicle, service or instrument.

DUSD

Over the past two weeks, the DUSD Machine, which executes Meccanico’s proprietary USD-denominated strategy on the Makina Protocol, generated a 5.15% APR on Meccanico’s deployed capital, with yield driven primarily by looped positions, alongside meaningful contributions from Sentora’s Morpho vaults, Convex crvUSD pools, and tokenized receivables (3Jane).

A key focus during the period was broadening the Machine’s underlying exposure. DUSD is pivoting away from concentrated looping strategies and towards real-world-asset (RWA) sources. This month the DUSD Machine added exposure to tokenized receivables through 3Jane’s senior tranche, USD3, tokenized HELOCs, which DUSD lends against on Morpho, and tokenized reinsurance through Re’s senior tranche, reUSD. Currently anticipated additions include mGLOBAL, an asset-backed lending tokenized fund.

Looking ahead, the most consequential currently-anticipated change will be the full wind-down of looped positions over the next one to two weeks.The DUSD Machine’s strategy is being deliberately concentrated into RWA exposures to reduce the Machine’s sensitivity to net interest margin during borrow-rate dislocations. We also anticipate adding exposure to FalconX’s CLO over the coming period.

Overall, the DUSD Machine’s recent adjustments reflect a continued emphasis on reducing rate-sensitive leverage and broadening yield generation through real-world assets and uncorrelated sources, while maintaining a highly liquid profile on the Machine level.

DUSD Summary:

DUSD Performance:

DUSD Asset Allocation Breakdown:

DUSD Protocol Allocation Breakdown:

DUSD Chain Allocation Breakdown:

DUSD Exposure:

DETH

Yield generation in ETH has become increasingly difficult over the past year. As staking yields compress and many of the easy opportunities disappear, managers are often pushed toward greater complexity in search of incremental returns. We believe the opposite approach is the right one.

Over the past two weeks, the DETH Machine, which executes Meccanico’s proprietary ETH strategy on the Makina Protocol, has generated an annualized yield of approximately 4% on Meccanico’s deployed capital. While that number may not seem spectacular in isolation, it highlights what we believe is the most important characteristic of the strategy at this stage: simplicity.

Rather than relying on aggressive leverage, opaque structures, or highly tactical trades, the DETH Machine has focused on maintaining a clean and resilient strategy construction. We characterize the result as a strategy that can once again generate yields that are safer, more scalable, and better predictable while preserving liquidity and keeping risk tightly controlled.

This period has been particularly encouraging because it demonstrates that sustainable returns can still be achieved while staying true to the core principles that have always guided the DETH Machine: capital preservation first, disciplined risk management, and a strong preference for opportunities that can scale.

That said, simplicity does not mean standing still.

Over the coming weeks, we anticipate rolling out a number of enhancements designed to expand the strategy’s opportunity set while remaining consistent with its risk framework. As a small preview, we are currently finalizing an ETH-focused arbitrage engine that will become one of the destinations for a portion of DETH’s idle liquidity. The objective is straightforward: capture market-neutral opportunities that can enhance returns without materially increasing directional exposure.

In parallel, another portion of the liquidity currently allocated to Lido is anticipated to be repurposed as collateral for an allocation into the DUSD Machine, our flagship stablecoin strategy. This integration is particularly attractive because it allows the DETH Machine to benefit from a broader set of yield sources while maintaining a conservative overall profile.

The intended combined effect of these changes should be twofold. First, they are anticipated to improve capital efficiency by reducing cash drag at the Machine level. Second, they are anticipated to increase diversification by adding return streams that are less dependent on any single source of ETH yield.

As always, our objective is to improve returns through thoughtful capital allocation rather than additional risk. The upcoming changes at the Machine level are incremental by design, but we believe they should contribute meaningfully to overall portfolio efficiency and help maintain a more consistent yield profile as the strategy continues to scale.

DETH Summary:

DETH Performance:

DETH Asset Allocation Breakdown:

DETH Protocol Allocation Breakdown:

DETH Chain Allocation Breakdown:

DETH Exposure:

DBIT

Over the past period, the DBIT Machine, which executes Meccanico’s proprietary BTC-related strategy on the Makina Protocol, has experienced a flat month due to negative returns in the first week for June in the DUSD Machine: a borrow rate spike on loops caused the underlying DUSD strategy to be slightly negative for a couple of days exacerbating the borrow cost for the DBIT Machine. DBIT allocates 35% of its Machine-level positions to DUSD and hence suffered.

DBIT remains small in scale with a few concentrated Machine-level allocations: Makina DUSD, AAVE (approximately 25% of the Machine’s position) - the borrow venue for the carry trade, and Across LP (40% of the Machine’s position). We anticipate to reduce the DBIT Machine’s exposure to DUSD by 50% while it is unwinding the loops which are a major cause of intra-day return volatility.

DBIT Summary:

DBIT Performance:

DBIT Asset Allocation Breakdown:

DBIT Protocol Allocation Breakdown:

DBIT Chain Allocation Breakdown:

DBIT Exposure:

Thanks for reading.

To learn more about Dialectic.

For the latest news, follow us on X.

DUSD, DETH, and DBIT Machines are being operated on Makina - the DeFi execution engine.

Have questions or comments? Head over to Discord and open a ticket or you can also find us in Makina’s Telegram.

Risk and disclaimer

Each Machine Token(DETH, DBIT, DUSD) is a product sold by Dialectic Meccanico Ltd. (BVI) as principal seller, acting for its own account. Each Machine Token is issued at the protocol level via the Makina Caliber smart contract on the Makina Protocol when the purchase price is paid. Machine Tokens are sold on an “AS IS, WITH ALL FAULTS” basis; users bear full risk, including total loss of the amount paid. Dialectic Meccanico Ltd. (BVI) operates the underlying Machines using its own capital. Machine Token holders participate at the protocol level; Meccanico does not manage assets for, or act on behalf of, Machine Token holders.

All Machine activity carries risks including smart-contract risk, oracle risk, market risk, liquidity risk, and protocol-governance risk. Past strategy performance is not indicative of future strategy performance. Investments involve risk, including potential loss of principal.

Nothing in this update should be considered investment, tax or legal advice or the recommendation to sell, or the offer of a solicitation to buy or invest in any investment product, vehicle, service or instrument.

SpaceX is the most important company of this century as its combined ambitions address the largest TAM of any company in history, far beyond the S1 estimation of 28.5T. SpaceX is growing at an unprecedented rate for a trillion-dollar tech conglomerate and is rapidly converging an array of highly synergistic businesses in its quest to make consciousness multi-planetary, deliver superior connectivity infrastructure and accelerate mankind via a maximally truth-seeking generalized artificial intelligence. Dialectic herein posits a near-term valuation target of $179/share (blended equity value $2.405T) with a long-term valuation target (10yr+) of 10T. Our first-principles, scenario-weighted, sum-of-the-parts DCF across the estimated twelve business units produces a P50 enterprise value of $166/share ($2.241 trillion), with meaningful right-skew to $352/share $4.733 trillion at P90. We will continue to accumulate SPCX up to our P55 of $179/share, continue opportunistically adding via proprietary strategies up to our P75 of $258/share and favour holding while employing our advanced yielding strategies on top of the asset there above until more information becomes clear. Our probabilistic sum of the parts DCF is complemented by a comparables analysis that considers a moderate ‘Elon Premium’ against peers, as well as current market signals.

We note that there is considerable upside potential from this base given the dominant market position in the various emerging markets that are tangential to space exploration, AI, and connectivity. We observe accelerating network effects across the approximately twelve (12) business lines currently operated under the SpaceX brand. While this document does not contemplate a pending highly strategic and synergistic merger with Tesla, we do note additional upside potential therein. At Dialectic, one of our principles is that ‘the future is far more Sci-Fi than you think’. SpaceX is pulling mankind towards a literal Sci-Fi future, more than any other company on the planet (or off planet) and we consider exposure to this asset fundamental for our future-forward investment portfolio.

Investment Thesis

SpaceX is rapidly accelerating its strategic advantages across the different business units in the primary categories 3 categories of 1- Space exploration, 2- connectivity, and 3- AI + Digital, with special consideration to the synergistic convergence of these businesses and resultant network effects.

Pillar 1: Space -- Network Effects & Moat in Access to Orbit

SpaceX has fundamentally repriced access to orbit, reducing commercial launch costs by over 90% since Falcon 9’s introduction. With 140-145 Falcon 9 launches expected in 2026 and Starship approaching commercial readiness in the later part of the year, SpaceX holds an incredibly advantageous competitive position and accounts for a significant majority of all launches globally by count and by weight-to-orbit. No competitor can match SpaceX on cadence, reliability, and price as Starship’s sheer scale drops the cost-to-orbit below $100/kg, driving under $10/kg by 2030. We are witnessing the SpaceX transcend Wright’s Law and push us towards a Space Age version of Moore’s Law. A fully operational Starship opens a $150B+ cargo-as-a-service near-term TAM.

Pillar 2: Connectivity -- Starlink as a Global Telecom & Internet Infra + D2C

Starlink is not just a satellite internet company. It is a global telecommunications network that will pass terrestrial communications infrastructure in cost, reliability and eventually latency/performance. It is an internet infrastructure business with recurring subscription-based revenue, low marginal cost per additional subscriber, and structural advantages over every terrestrial and geostationary competitor. At 12+ million subscribers across 164 countries and growing over 100% annual CAGR, Starlink generates $11.4 billion of annual revenue with a 63% EBITDA margin. Direct-to-Cell extends the Starlink franchise to mobile devices globally and should capture some of the approximately $400 billion in cellular telecom annual profits.

Pillar 3: Maximally Truth-seeking AI -- The AI Infrastructure Stack

The winner in AI Foundation Models - which we often refer to within Dialectic as “the Sport of Kings” (Bill Gurley, 2024) for its extraordinary capex vs TAM dynamic - will be decided on 3 fronts: 1- Data, 2- Compute, and 3- Algorithmic innovations. The 2025 merger with xAI created a unique vertically integrated AI company: the only entity that simultaneously trains frontier AI models (Grok), operates its own GPU cluster at unprecedented scale and speed of execution (Colossus), distributes models to 600+ million users (X.com) which generates exclusive proprietary training data and financial transactions (Xmoney), while building orbital compute infrastructure supported by fabrication of semiconductor chips for both terrestrial edge-nodes and orbital training, all connected via its own global communications network.

An important but misunderstood advantage of Grok/xAI is the principle to develop a maximally truth-seeking AI -- this allows Grok to learn and develop at a different pace than models that have an ideological throttle placed on them. A maximally truth-seeking AI also builds deep trust with users and will ultimately deliver superior performance in recursive learning environments.

Colossus I recently closed $26 billion in agreements with Anthropic and Google demonstrating significant growth potential of Colossus in the short term, while SpaceX prepares for the future of Orbital Data Centers.

Pillar 4 -- The Convergence Flywheel

It will become more obvious over time that the true value of SpaceX is not merely the sum of its parts -- it is their synergistic intersection and resultant growth. SpaceX manufactures satellites, launches them on its own rockets, and produces the full global connectivity stack, connecting with its own manufactured edge-nodes with Tesla robots and potential internally developed edge-node devices. X’s 600M+ users generate exclusive training data for Grok, a maximally truth-seeking AI. XMoney (launched April 2026 with Visa integration across 41 US states) monetizes those interactions with embedded fintech. Starshield and Golden Dome contracts ($6.45 billion awarded in 2026 alone) leverage the same orbital expertise built for Starlink as the defining ‘defense neo-prime’ in a time of significant geopolitical uncertainty. ODCs present medium-term advantages with respect to energy, cooling, networking, and permitting. These businesses support off-planet resource extraction, manufacturing and eventually interplanetary colonization. We often ask ourselves, ‘what is the TAM of the boundless potential of space?’. Every business unit strengthens every other and is unified under a singularly ambitious mission and common culture often referred to “as a zone of shocking competence” (Marc Andreesen, 2026), where the pace of work and growth is unmatched globally.

Business Unit Analysis

SpaceX is a somewhat complex conglomerate operating across a range of current and near-term. For the sake of simplicity in this first iteration of our analysis, we have divided the business into 12 segments, worked through revenue and profitability potential on 3 different horizons, while heavily weighting near-term potential for the sake of conservatism. By some internal accounts SpaceX is actually more than two dozen different businesses and we built up revenue projections by looking at each one from first principles, using public sources. We expect public figures to become more widely known post the 2026-Q2 reporting and intend to update this analysis as new information becomes available. This is an incredibly dynamic business.

1 -- Launch Services

The Orbital launch segment generated $4.086 billion of revenue in FY2025 at roughly $653 million of adjusted EBITDA — a ~16% margin that has been depressed during the Starship development cycle. SpaceX accounts for roughly 85% of all global satellite launches, having repriced access to orbit; we expect 140-150 Falcon 9 launches this year at roughly $74 million per launch. Our model update revises the 2026 base revenue to $11.5 billion. The bull case requires some Starship commercial revenue to materialize in 2026-2027. Given competitor challenges in recent months, this seems. We accounted for lower EBITDA margins in the near term by Starship R&D (~$15B). As Starship becomes operational and the development drag rolls off, margins should converge toward 40-50%. Our scenarios apply revenue CAGRs of roughly 15% / 25% / 45% (Bear / Base / Bull), and S-1 calibration lifted our base-case revenue anchor from ~$6.8B to ~$13.5B as the HLS Artemis lunar-lander and Starship cadence ramp.

2 -- Crewed Trips, 3 -- Other Space Services

NASA CCP/CRS and deposits received for manned private spaceflight make up current revenues. NASA crewed missions account for $700M-$900M of annual revenue inside of this decade, supplemented by Polaris and Axiom private-astronaut missions. There are distinct plans for private spaceflight, both interplanetary and globally (global point-to-point in less than 45 minutes via Starport infrastructure) that could be significant as well. Other space services — satellite servicing, orbit-to-orbit transport (orbital tug), and rocketry IP licensing or manufacture for sale — are a smaller line modeled conservatively but capable of opening entirely new business lines. These segments are also where SpaceX’s Mars ambitions begin to monetize: as the cadence and payload economics improve, the path toward a self-sustaining off-planet presence underwrites optionality that the comps cannot capture. Elon has referenced a space hotel in a recent JP Morgan roadshow appearance. Overall, we have these segments as relatively benign in the scenario weighting, but we maintain them as a placeholder as there is outlier potential for these segments over time, particularly as the core mission of SpaceX (and Elon’s compensation) are centered around a threshold of inhabitants on Mars.

4 -- Starlink Connectivity

Starlink is SpaceX’s cash cow: $11.39B revenue and roughly $7.17B of adjusted EBITDA in 2025, a 63% margin that expanded from ~50% as the constellation scaled and fixed costs leveraged. 12M+ subscribers across 164 countries. ARPU compressed from $86 to $66 as Starlink expanded into lower-income geographies, but the May 2026 price increase signals a deliberate shift from land-grab toward monetizing the existing base. Aviation and maritime adoption are rapidly accelerating and carries materially higher ARPU than the consumer tier. S-1 calibration lifted our base-case revenue anchor from ~$9.8B to ~$18B, and our scenarios apply revenue CAGRs of roughly 15% / 35% / 55% (Bear / Base / Bull) with margins trending toward the mid-60s.

5 -- Starlink Direct-to-Cell

DTC represents an entirely new revenue stream, monetized initially through carrier licensing fees from T-Mobile, KDDI, Optus, Rogers and others. The addressable opportunity is large — global cellular profits run ~$400B/yr, and the directly serviceable supplemental-coverage market is on the order of $40-100B — of which Starlink should be able to capture a significant component. Crucially, Direct-to-cell rides on the Gen2 constellation, which is already launching, so the marginal cost of per additional client is relatively low. Bilateral spectrum agreements now exist in dozens of countries. This is the most binary segment in the model on a 5-year basis: outcomes hinge on spectrum, handset compatibility, regulatory/political questions, and carrier deal-structure rather than on technical feasibility - which is now demonstrated. With the new generation of Starlink satellites offering superior speeds, we could very well see massive adoption growth in home and businesses as Starlink goes from internet back-up or remote use-cases to core home and business internet connection.

6 -- Starshield & Golden Dome

SpaceX is the best-positioned defense neo-prime contractor in an era of considerable geopolitical uncertainty. $6.45B in new Golden Dome contracts in 2026: the SDN Backbone at $2.29B (with an end-2027 prototype milestone) plus AMTI satellites at $4.16B. Additional awards: an NRO $1.8B classified satellite contract and a Space Force PLEO IDIQ ceiling of $13B. In aggregate, SpaceX’s recent defense awards exceed the combined awards of all other companies in the segment, underscoring its leading neo-prime status. These contracts are multi-year, so initial revenues are conservatively projected with lower weighting; our scenarios run a base case near $2.3B scaling toward a bull case of roughly $6.5B as Golden Dome and Starshield production ramp.

7 -- xAI Software (Grok + API)

Grok + API $3.201B of revenue in FY2025 against an operating loss of roughly $(6.355B) — the heavy training-and-talent investment that funds Grok’s position among the top two-to-three frontier models. The application layer is inflecting fast: Cursor went from $1B in ARR in November to $2B+ in ARR in March, and we assume the ~$60B acquisition is likely to consummate shortly after the IPO. Monetization runs across SuperGrok, enterprise API, and rapidly growing DoD / Department-of-War contracts, with X distribution (600M+ users) providing a low-cost-CAC funnel. AI Foundation models, or “the Sport of Kings” is an incredibly competitive and capital-intensive indusrty and thus the AI business is currently unprofitable. We expect inflection to profitability by late 2026 into early 2027 as inference costs fall and enterprise/API revenue scales against a largely fixed base. Also, Macrohard could prove very compelling and Grok Build is rapidly scaling, entering an era of recursive self-improvement and asymptotic intelligence.

8 -- Colossus Terrestrial Data Centers

The 300MW, 220,000 NVIDIA GPU-equivalent cluster in Memphis — stood up in a now-famous 122-day build that demonstrated SpaceX/xAI’s shocking speed advantage in deploying compute. The $15 billion Anthropic and $11 billion Google agreements validate the market-clearing price for frontier GPU-hours and their resultant tokens, and these contrasts de-risk the buildout by underwriting third-party demand alongside xAI’s own training needs. SpaceX targets roughly 1 million GPU-equivalent additions per year, a commitment it intends to sustain across a multi-year horizon. Colossus II is now operational, training novel Grok models, Macrohard, and it serves as the terrestrial bridge to the Orbital Data Center roadmap.

9 -- X Platform & XMoney

550-600M MAU and roughly 245-250M DAU, with $3.5B of FY2025 revenue of which ~70-75% is advertising — a base that has recovered strongly from the ~$2.7B 2023 trough. XMoney launched April 2026 with Visa integration, P2P payments across 41 US states, >100M DAU on the payments surface, and 6% APY on deposits. Our model raises X Platform’s 2026 base revenue to $7B, reflecting XMoney, creator monetization, AI-targeted advertising powered by Grok, and an expanding app scope toward the WeChat-style super-app model — a transition that, if it lands, drives non-linear revenue per user rather than the linear ad-load economics the current multiple assumes.

10 -- Orbital Data Centers

In-space compute with distinct advantage – lowered levelized cost of energy with high capacity-factor solar arrays, improved cooling through highly engineered radiators, faster networking via lasers through the vacuum of space, and the elimination of the NIMBY/permitting constraint (holding up nearly 50% of planned US data centers). This business line is conservatively classified as speculative; estimated first commercial operation in 2030, noting that there is a reasonable chance that ODCs come online before that and have significant upside similar to Colossus or are potentially delayed in achieving scale. The transformational case is cost-per-FLOP parity with terrestrial compute: once orbital launch and on-orbit assembly costs fall far enough that a watt of space-based compute is competitive with a watt on the ground, the structural advantages above (energy, cooling, networking, no permitting) convert a niche into a potential platform shift in where large-scale AI training is done — a market SpaceX is uniquely positioned to serve given its launch, satellite, and chip-fabrication stack.

11 -- Deep Space Optionality

Deep Space Optionality includes a range of far-reaching businesses that are on the roadmap, but possess economic models that are not as well understood — Lunar then Mars colonization, resource extraction on other planets and asteroids, orbit-to-orbit transport, and exploratory R&D. We note that ambitious targets, such as 2028 lunar missions, sustain a narrative premium and attracts the best talent in aerospace. Our model treats this as a small probability-weighted DCF contribution that we treat as a deep-DCF, so it does not double-count optionality recognized elsewhere.

12 -- Terafab

The vertically integrated semiconductor fabrication project announced in March 2026 building 1- custom space-optimized ASICs for training in space, 2- high-performance compute for Colussus, and 3- specific edge-node inference chips both for Tesla robots and cars alongside a potential edge- node device with telecommunications (think, the next generation of the smart phone) that SpaceX is uniquely positioned to offer. The world’s largest chip-manufacturing facility under one roof, and the most vertically integrated: Terafab is designed to consolidate chip design, lithography, HBM memory, advanced packaging, and test under a single roof — the steps that today are fragmented across TSMC, ASML, SK Hynix and others. It would supply SpaceX’s own roadmap (AI5-class FSD silicon, AI6-class Optimus inference, and D3-class space-hardened training chips) rather than depend on the merchant market. Target capacity is over 1 terawatt of AI compute per year. First phase is ~$55B, with full buildout up to $119B. Revenue is modeled beginning 2031 to account for the unknown unknowns in chip design and foundry development, making this the longest-dated and most heavily discounted line in the SOTP.

Valuation Analysis

Methodology

We primarily employed a bottoms-up, probability-weighted sum-of-the-parts (SOTP) DCF across twelve business units, each modeled with three scenarios (Bear / Base / Bull) and a materialization probability considered in short, medium and long time horizons with strong overweighting near-term cash flows. The DCF tab discounts the resulting UFCF at a blended WACC of 9.0% with a terminal growth rate of 4.0%. In addition to this valuation model, we evaluate comparables in relevant industries and consider the Elon Premium alongside Current Market Signals.

Comparables Analysis

We treat trading comparables as a cross-check and a downside bound, not as the primary basis for value. SpaceX has no clean peer. A blended whole-company multiple is therefore directional at best, which is why our central value rests on the sum-of-the-parts DCF.

Viewed across the relevant peer sets, the comparables span an extremely wide band. Defense primes trade at roughly 2x revenue and anchor low, while high-growth satellite and new space names (Rocket Lab, AST SpaceMobile) trade at multiples that, applied to SpaceX’s revenue base, would imply values up to ~$2.47 trillion. The AI cohort sits between these poles: OpenAI screens at roughly 28x Revenue and Anthropic at roughly 21x, Palantir — the closest listed AI/defense neo-prime analogue — at roughly 47x Revenues, alongside NVIDIA at roughly 22x. Applying the AI and high-growth multiples to SpaceX’s revenue implies an equity value in the $1.0-1.7 trillion zone, broadly consistent with the DCF.

The synthesis is that comparables price the company as it is today, not the convergence flywheel we underwrite, and they penalize SpaceX for being mid-cycle in a heavy-investment phase. They are therefore most useful as a soft floor — framing value near $0.9-1.0 trillion — with the DCF and the IPO range describing where the market is actually likely to clear, given the Elon Premium.

Elon Premium

Across peer groups, SpaceX valuation implies an average Elon Premium of roughly 1.8x against the AI-lab median of OpenAI and Anthropic (24.5x EV/Revenue), while being in line with Palantir, the closest listed AI/defence neo-prime comparable at 46.7x EV/Revenue vs SpaceX’s 43.5x EV/Revenue and 79.1x EV/EBITDA (2026E, DCF conclusion) to frame the premium. On EBITDA, the picture inverts the headline: SpaceX screens roughly 80% Palantir’s 100.0x, i.e. the one preferred peer that is profitable. OpenAI and Anthropic carry no EBITDA multiple as both remain pre-EBITDA. Surprisingly, our valuation has SpaceX at the same revenue multiple as space industry peers despite extraordinary structural advantages.

The Elon Premium is best understood as compensation for the compounding optionality of an operator who moves faster, builds more ambitiously and competently, and controls more of the value chain than any single peer. SpaceX has no clean comparable.

Current Market Signals

Hyperliquid and Polymarket have been pricing the enterprise valuation between $2.17 and $2.45T for the last several weeks with $42M and $2.3M in daily volume respectively, which substantiates the legitimate market depth of these positions.

Risks & Key Considerations

Execution & Technology Risk: Starship’s commercial timeline is a risk, as is the timeline of direct-to-cell, Terafab and to a lesser extent Colossus, and Starshield. Numerous projects, in particular Terafab, represent enormous capital commitments with binary outcomes. If the $55 billion Terafab Capex does not produce cost-competitive chips for AI training, it could be written down.

Key-Man Risk: All of the businesses are dependent on Elon Musk for design direction and important cultural leadership. Any event that reduces Musk’s involvement -- legal, political, or personal -- would likely cause rapid multiple compression, even if the risk event doesn’t affect business fundamentals or competitive positioning.

Regulatory & Political Concentration Risk: SpaceX’s Starshield and Golden Dome revenues are politically concentrated in the current administration. Particularly, an antagonistic Democrat-controlled branch of government could hold back the business in different ways. A change in administration could materially slow or cancel government programs. Other countries and companies could decide not to work with SpaceX for geopolitical reasons, thereby propping up the competition and digesting some market potential over time. Regulatory risk is also the primary concern with respect to a potential merger with Tesla.

Competition & Market Structure: Our analysis concluded that the most significant competition across the business units of SpaceX comes primarily from China today, both from state-sponsored projects as well as private enterprises such as i-Space and LandSpace. It would be unwise to underestimate the Chinese in their ability to develop, manufacture, and push the boundaries of exploration. Space pure-play competitors in America do exist and are growing with formidable backing – Amazon’s Blue Origin, Rocket Lab ($RKLB), and Eric Schmidt’s Relativity are all relevant, even amidst some technical setbacks and scale limitations therein.

Note: We observe with opportunistic excitement the flowering of an array of ‘SpaceX Cubs’ or ‘SpaceX Mafia’ – offspring led by former SpaceX team members that are building compelling technologies. Impulse, Verda, Apex, Firefly and others may present new opportunities for acquisition and these different experiments may prove positive.

Finally, but very importantly, the cost of serving intelligence in AI is rapidly decreasing, which could have profound impacts on society and industry. The nature of AI may fundamentally change to primarily edge-computing, open-source. Recursive learning could result in totally unforeseen solutions to the compute dirth. There are many ways AI could commodify quickly, rendering massive previous capital expenditure as sunk cost.

Financial & Balance Sheet Risk: Total long-term debt of $29.1 billion against cash of $15.9 billion produces net debt of $13.2 billion. Q1 2026 burned $8.9 billion in cash (R&D and xAI integration costs). The $75 billion IPO raise is indeed necessary for short term growth. Further dilutive fundraising could drag down return profile in the short and medium term.

The Elon Premium: The Elon Premium is real, measurable, and -- in our view -- historically justified. SpaceX at 43x 2026E Revenue requires the market to continue believing in the full convergence flywheel and for multiple business units to have successful outcomes. Note that at $1.5 trillion ($112/share), the Elon Premium comes for free.

Conclusion

When we look up towards the stars we are reminded of the infinite possibilities of space and the inspiring Sci-Fi future that mankind will chart as a multi-planetary species. SpaceX is uniquely positioned as the most consequential private company of the 21st century, potentially in the history of capitalism. The flywheel of vertical integration across these businesses is compounding in ways that no competitor can replicate. At 43.5x 26E Revenues ($55.6B) and our DCF blended fair value of $179/share ($2.418T), we reiterate our conviction on the accumulation plan set herein. The opportune issuance of SPCX may be remembered as the most extraordinary gift to America on the eve of her 250th Anniversary. At current levels the investor buys the current business across launch, connectivity, AI and digital. The Moon, Mars and potentially many other celestial bodies are free, as is the impact of a maximally truth-seeking generalized intelligence as recursive self-improvement takes hold.

ANNEX A | PROBABILITY-WEIGHTED VALUATION SUMMARY

A

Probability-Weighted Valuation Summary

Blended EV, equity value, and per-share across 5yr, 10yr, and 20yr horizons. WACC 9.0%, terminal g 4.0%, net debt $13.3B, 13,400M diluted shares.

Blended weights: 5yr 55%, 10yr 35%, 20yr 10%. Horizon weights set by investment committee to reflect near-term certainty premium.

Figure A1: Enterprise value by horizon. Blended (55/35/10 weighted) = $2,421B.

Monte Carlo Distribution (5,000 Trials)

Figure A2: Monte Carlo equity distribution. P10 shaded red, P90 shaded green.

ANNEX B | SUM-OF-THE-PARTS VALUATION

B

Sum-of-the-Parts Valuation

Per-segment PV of 10-year UFCF + Gordon Growth terminal value. WACC 9.0%, terminal g 4.0%.

WACC 9.0% | Terminal g 4.0% | PV UFCF = SUMPRODUCT of 10yr UFCF with discount factors | PV TV = Gordon Growth on year-10 UFCF.

Figure B1: SOTP enterprise value by segment. Bear / Base / Bull shown. Starlink Connectivity and Starlink Direct-to-Cell are the two largest contributors at 33.0% and 16.8% respectively.

ANNEX C | CONSOLIDATED REVENUE & UFCF PROJECTIONS

C

Consolidated Revenue & UFCF Projections

Probability-weighted consolidated P&L by business unit, 2026-2035. All figures in $M.

Revenue CAGR 2026-2035: ~34%. UFCF CAGR: ~37%.

Figure C1: Consolidated revenue by business unit, stacked 2026-2035.

Figure C2: UFCF (red, left axis) and total revenue (dark, right axis) 2026-2035.

DCF Build by Horizon

ANNEX D | COMPARABLE COMPANIES ANALYSIS

D

Comparable Companies Analysis

Trading multiples. SpaceX 2026E: revenue $55.6B, EBITDA $30.6B.

Figure D1: EV/Revenue multiples. SpaceX at DCF conclusion (43.5x) shown in red.

ANNEX E | ELON PREMIUM & IMPLIED VALUATION METHODS

E

Elon Premium & Implied Valuation Methods

Peer-multiple implied valuations applied to SpaceX 2026E financials ($55.6B revenue, $30.6B EBITDA).

ANNEX F | SENSITIVITY & SCENARIO ANALYSIS

F

Sensitivity & Scenario Analysis

WACC x terminal growth grid and sensitivity tornado. Base: WACC 9.0%, g 4.0%, EV $2,421B, equity $2,408B.

Table F1 -- Blended Enterprise Value ($B): WACC x Terminal Growth Rate

Table F2 -- Value per Share ($): WACC x Terminal Growth Rate

Figure F1: Equity value heat map. Base case (WACC 9.0%, g 4.0%) outlined in dark.

Table F3 -- Sensitivity Tornado

Figure F2: Sensitivity tornado. Drivers sorted by range. WACC and governance discount are the dominant risks.

ANNEX G | BUSINESS UNIT DETAIL -- REVENUE, EBITDA & FCF

G

Business Unit Detail -- Revenue, EBITDA & FCF

Probability-weighted revenue, EBITDA, and UFCF for each of the 12 segments, 2026-2035.

1 | ANCHOR | Base EV $324B

Launch Services & Lunar Vehicles

Falcon 9/Heavy commercial + Starship cargo ramp. S-1 Space segment 2025: $4.1B. CAGR 2026-2035: ~33%.

Figure G1: Revenue & EBITDA (left), UFCF (right), 2026-2035.

2 | ANCHOR | Base EV $63B

Crewed Spaceflight

NASA CCP/CRS Dragon + Axiom, Polaris, private spaceflight. CAGR 2026-2035: ~23%.

Figure G2: Revenue & EBITDA (left), UFCF (right), 2026-2035.

3 | PLAUSIBLE | Base EV $6B

Other Space Services

Satellite servicing, orbital tug, rocketry IP licensing. CAGR 2026-2035: ~31%.

Figure G3: Revenue & EBITDA (left), UFCF (right), 2026-2035.

4 | ANCHOR | Base EV $745B

Starlink Connectivity

10M+ subscribers, 164 countries. S-1 2025: $11.4B, 62.9% EBITDA margin. CAGR 2026-2035: ~32%.

Figure G4: Revenue & EBITDA (left), UFCF (right), 2026-2035.

5 | PLAUSIBLE | Base EV $380B

Starlink Direct-to-Cell

~300 DTC satellites. T-Mobile, KDDI, Optus, Rogers. CAGR 2026-2035: ~62%.

Figure G5: Revenue & EBITDA (left), UFCF (right), 2026-2035.

6 | ANCHOR | Base EV $190B

Starshield & Golden Dome

$6.45B in new Golden Dome contracts in 2026. NRO + PLEO IDIQ $13B ceiling. CAGR 2026-2035: ~35%.

Figure G6: Revenue & EBITDA (left), UFCF (right), 2026-2035.

7 | ANCHOR | Base EV $86B

xAI Software (Grok + API)

Top-3 foundation model. Maximally truth-seeking. Cursor ~$60B acquisition likely. CAGR 2026-2035: ~40%.

Figure G7: Revenue & EBITDA (left), UFCF (right), 2026-2035.

8 | ANCHOR | Base EV $329B

Colossus Data Centers

220,000 NVIDIA GPUs. $15B Anthropic agreement. 1M GPU-equivalent additions/year. CAGR 2026-2035: ~30%.

Figure G8: Revenue & EBITDA (left), UFCF (right), 2026-2035.

9 | PLAUSIBLE | Base EV $53B

X Platform & XMoney

600M+ MAU. XMoney: Visa, 41 states, 6% APY. 2026 base: $4.1B. CAGR 2026-2035: ~27%.

Figure G9: Revenue & EBITDA (left), UFCF (right), 2026-2035.

10 | SPECULATIVE | Base EV $27B

Orbital Data Centers

In-space compute: perpetual solar, deep-space radiative cooling, no NIMBY. First commercial 2031.

Figure G10: Revenue & EBITDA (left), UFCF (right), 2026-2035.

11 | SPECULATIVE | Base EV $0.02B

Deep Space Optionality

Moon/Mars cargo 2028, asteroid extraction, P2P transport. Narrative premium and talent magnet.

Figure G11: Revenue & EBITDA (left), UFCF (right), 2026-2035.

12 | PLAUSIBLE | Base EV $54B

Terafab -- Semiconductor Mfg

World largest chip fab. AI5/AI6 + D3. $55B first phase. Revenue modelled from 2031.

Figure G12: Revenue & EBITDA (left), UFCF (right), 2026-2035.

DUSD

Over the past two weeks, the DUSD Machine generated a 5.89% APR while Dialectic Meccanico continued to refine the Machine’s liquidity profile and diversify underlying exposure.

A key change during the period was the reduction of illiquidity in the strategy. Exposure to InfiniFi moved from a four week lock to a one week lock, with the intention to further reduce concentration over the coming weeks. This will allow the Machine to achieve greater granularity across real world asset exposures. Currently anticipated additions include Re, focused on reinsurance, and USD3 by 3Jane, focused on asset-backed finance.

The Machine also introduced limited Royco’s junior tranche exposure through staked Cap and syrupUSDC. These positions have remained capped at no more than 3% of AUM of the Machine’s allocations over recent weeks. Meccanico characterizes these allocations as a capital efficient form of embedded leverage, without introducing direct interest rate risk.

Overall, the DUSD Machine’s recent adjustments reflect a continued emphasis on reducing leveraged exposures sensitive to interest rate risk, while broadening yield generation through real world assets.

DUSD Summary:

DUSD Performance:

DUSD Asset Allocation Breakdown:

DUSD Protocol Allocation Breakdown:

DUSD Chain Allocation Breakdown:

DUSD Exposure:

DETH

The DETH Machine’s strategy remained resilient over the past period, with the current positioning continuing to perform as expected. Borrow rates across the relevant venues have remained relatively stable, which has helped preserve a positive spread for the looping strategies currently deployed. This has allowed the portfolio to maintain consistent carry without requiring meaningful changes to the underlying allocation.

Looking ahead, Dialectic Meccanico is actively exploring an integration with the DUSD Machine as a way to further expand the subset of available carry trade opportunities within the strategy. Given DUSD’s role as one of the Machines Dialectic Meccanico operates, we see this as a natural venue to broaden the strategy set and introduce additional sources of yield where the risk/reward profile is appropriate. This integration is currently anticipated, subject to operational and market conditions.

More specifically, delegating the stablecoin leg to the DUSD Machine feels like a natural extension of the current strategy design. Since Meccanico is structuring these strategies in a coherent and coordinated way, this allows the different Machines to specialize in, and delegate exposure to, the subset of a strategy that is most aligned with their operational focus. In this case, the DETH Machine can retain the ETH-denominated exposure and core looping logic, while the DUSD Machine can act as the dedicated stablecoin-side vehicle supporting the carry trade.

Beyond this, Dialectic Meccanico continues to monitor new integrations across ETH-based lending and liquidity provisioning markets. The focus remains on opportunities that can either improve the current yield profile or add complementary sources of return without materially changing the overall risk framework of the strategy. Overall, the DETH Machine’s positions remained stable during the period.

DETH Summary:

DETH Performance:

DETH Asset Allocation Breakdown:

DETH Protocol Allocation Breakdown:

DETH Chain Allocation Breakdown:

DETH Exposure:

DBIT

In the second half of May, the DBIT Machine experienced somewhat depressed returns of 0.18% APR due to cbBTC’s price volatility vs WBTC and borrow rate spikes. Meccanico has reduced all exposure to cbBTC to 0%. Following this adjustment, the strategy currently targets returns within the historical range of recent months (approximately 1-2%), subject to market conditions and venue-rate dynamics.

DBIT’s allocations remain concentrated in the Across LP pool, AAVE, and Makina DUSD. The previous exposure to Morpho has been reduced to 0% due to a borrow rate spike. The Machine’s position is fully liquid (the DUSD portion has 12h liquidity). No semi-liquid or non-liquid positions are currently anticipated for introduction in June

DBIT Summary:

DBIT Performance:

DBIT Asset Allocation Breakdown:

DBIT Protocol Allocation Breakdown:

DBIT Chain Allocation Breakdown:

DBIT Exposure:

Thanks for reading.

To learn more about Dialectic.

For the latest news, follow us on X.

DUSD, DETH, and DBIT Machines are being operated on Makina - the DeFi execution engine.

Have questions or comments? Head over to Discord and open a ticket or you can also find us in Makina’s Telegram.

Risk and disclaimer

Machine Tokens (DETH, DBIT, DUSD) are issued at the protocol level via the Makina Caliber smart contract. Dialectic Meccanico Ltd. (BVI) operates the underlying Machines using its own capital. Machine Token holders participate at the protocol level; Meccanico does not manage assets for, or act on behalf of, Machine Token holders.

All Machine activity carries risks including smart-contract risk, oracle risk, market risk, liquidity risk, and protocol-governance risk. Past performance is not indicative of future results. Investments involve risk, including potential loss of principal.

Nothing in this update should be considered investment, tax or legal advice or the recommendation to sell, or the offer of a solicitation to buy or invest in any investment product, vehicle, service or instrument.

DUSD

Over the past 14 days, DUSD delivered a 4.47% APR, supported primarily by looped positions on Morpho, with the largest contribution coming from the Maple, Reservoir, alongside additional exposure to InfiniFI four week lock.

Over the past few weeks, the machine navigated borrow rate spikes across lending venues following the Kelp incident and fully removed exposure to Umbrella. During this period, there were several unknowns around how the situation would evolve.

The borrow spike temporarily affected loops in Morpho, but it did not result in negative prints, as loops were reduced to a less aggressive health factor across positions. At the same time, the temporary spike in our Umbrella positions helped offset the impact, resulting in a net gain during the borrow spike period. Umbrella positions were exited on the 22th of April: tx hash.

Portfolio rotation during those weeks was primarily focused on capital preservation, given the broad market dynamics following the incident. During this period, we temporarily deposited GHO into the savings product offering 5.7%, while continuing to push new integrations to diversify the balance sheet away from a loop centric book and toward an uncorrelated source of yield (e.g: CLOs and ABL).

Over the coming weeks, we plan to roll out the following integrations in the root:

• Royco’s Junior tranche exposure: stcUSD and syrupUSDC

• FalconX CLO

• Exposure to naked mGLOBAL: asset backed lending tokenised fund managed by Fasanara

Overall, DUSD continued to outperform SOFR + 150 bps despite an eventful market environment. In parallel, the team has begun discussions around the potential use of DUSD as collateral across Morpho, while Dialectic is also conducting due diligence on select real-world asset opportunities for potential future integration, with the aim of offering a more blended source of yield while maintaining a highly liquid vault profile. We expect to share further updates on these initiatives in the coming weeks

DUSD Summary:

DUSD Performance:

DUSD Asset Allocation Breakdown:

DUSD Protocol Allocation Breakdown:

DUSD Chain Allocation Breakdown:

DUSD Exposure:

DETH

DETH has suspended conversion availability following the Kelp DAO incident, pending evaluation of any contingent exposure transmitted to the Umbrella position, 10% of book.

Based on the current outcome and the unstaking process already initiated, which is expected to complete on Sunday, 10 May, repurchases are expected to be re-enabled shortly after and the strategy should return to full operational status.

During the same period, the spike in ETH borrow rates caused by frozen Aave markets created a negative spread on the ETH looping positions, which had an impact on strategy PnL. To mitigate this, idle wstETH was used to borrow ETH more cheaply and reduce exposure on Aave.

Following the liquidation of the hacker on Core, borrow rates have started to normalize and spreads are moving back into positive territory. As seen several times in the past, these spikes tend to be temporary. Over a longer time horizon, ETH looping has continued to prove itself as one of the strongest scalable ETH strategies, and we remain confident in maintaining exposure to it. That said, allocation will be actively managed depending on overall book size, with a target exposure of approximately 30% at any given time.

Looking ahead, DETH will continue to diversify the opportunity set. The next areas of focus are selective exposure to stablecoin-related carry trades, as well as LP positions in incentive-driven pools across StakeDAO and Convex. These additions should help reduce reliance on a single source of ETH yield while improving capital efficiency and smoothing returns over time.

DETH Summary:

DETH Performance:

DETH Asset Allocation Breakdown:

DETH Protocol Allocation Breakdown:

DETH Chain Allocation Breakdown:

DETH Exposure:

DBIT

DBIT returns have normalized towards a stable return range over the past weeks as we reduced exposure to USDT. Currently, due to its small size, DBIT only has three allocations: Across Hubpool LP, borrow position on Morpho on Monad taking advantage of cheap borrow rates, and a Steakhouse Reservoir vault position on Morpho Ethereum. DBIT unwound all remaining illiquid positions (Umbrella) a few weeks ago and remains fully liquid. DBIT has exited its DUSD allocation to service conversion demand, with plans to reallocate in the coming weeks.

As noted above, because DBIT allocates capital to DUSD, Dialectic Meccanico will perform the necessary accounting to ensure that DBIT LPs are not charged twice. We will also publish a public repository so that the fee distributions and rebates associated with DBIT’s allocation to DUSD can be tracked transparently.

Overall, our outlook on DBIT remains positive despite the market volatility. With upcoming improvements to DUSD, we expect a further yield uptake.

DBIT Summary:

DBIT Performance:

DBIT Asset Allocation Breakdown:

DBIT Protocol Allocation Breakdown:

DBIT Chain Allocation Breakdown:

DBIT Exposure:

Thanks for reading.

To learn more about Dialectic.

For the latest news, follow us on X.

DUSD, DETH, and DBIT Machines are being operated on Makina - the DeFi execution engine.

Have questions or comments? Go to Discord and open a ticket or you can also find us in Makina’s Telegram.

Eight frontier AI models are competing to manage real DeFi portfolios with real capital.

The leaderboard is public and updates every eight hours, with every trade settling onchain.

Why Dialectic is Benchmarking Frontier Models in DeFi

Dialectic has been allocating capital in DeFi since 2019, delivering consistent risk-adjusted yield across every market regime this space has produced. Fully automated onchain execution went live in 2020. Institutional-grade risk management followed in 2022, including onchain insurance and preemptive risk protection. Along the way, the team assembled what has become one of the deepest proprietary datasets in DeFi, purpose-built for machine learning and analytics.

Over the last several months our team has developed specialized agent harnesses for specific tasks within the capital allocation process. We fine-tuned prompts and built custom MCP tools and skills to support each stage.

We have condensed over five years of DeFi operational knowledge into systems that cover the full cycle of risk-adjusted onchain yielding: scouting, due diligence, execution, harvesting, reporting, and audit. As of today, humans act in a checks-and-balances capacity, verifying outputs and refining the agents and their harnesses, with more automation every passing week.

This work raised a question.

If frontier AI models can now participate in onchain capital management, how well do they perform when given real capital in a live market?

Dialectic has the infrastructure and the domain expertise to run that test properly, and DeFi Bench is the demonstration.

The Setup

Eight models are competing in Season 1:

• @claudeai (@AnthropicAI)

• @ChatGPTapp (@OpenAI)

• @GeminiApp (@Google)

• @grok (@xai)

• GLM (@Zai_org)

• @Kimi_Moonshot (Moonshot)

• @MistralAI

• @Alibaba_Qwen (Alibaba)

All execution runs through Makina’s DeFi execution engine, with model inference hosted by Venice.AI. Each competitor receives identical seed prompts and the same starting capital in USDC, operating across the same approved protocols: @Morpho, @aave V3, @CurveFinance, and @0xfluid on @ethereum. The only variable is the model itself.

Each competitor runs as a group of three specialized agents, all powered by the same underlying model. It is crucial to understand each agent and their role:

• Risk Researcher: Runs before every round. Conducts due diligence on protocols and pools, researching security audits, exploit history, asset quality, and market dynamics. Produces a Risk Note with risk classifications and alerts that the Trader agent consumes. The Risk Researcher can reference its own previous assessments to track how conditions evolve over time.

• Trader: The decision maker. Receives the Risk Note, portfolio state, live market data, and any Investment Committee feedback. Produces allocation decisions and executable Makina commands (What’s a Makina command? More on this below). Maintains a persistent Strategy Document that records allocation targets, risk thresholds, and lessons learned across rounds.

• Investment Committee (IC): Runs every five rounds. Reviews performance and decision quality across both the Risk Researcher and the Trader. Evaluates whether the strategy is working, identifies patterns, and can directly update the Trader’s Strategy Document when changes are warranted. Produces an IC Assessment that feeds into subsequent rounds.

This separation mirrors how institutional investment teams operate where research informs trading, and periodic oversight keeps both accountable. The group of agents runs identically for every competitor.

The seed prompts are published on the official DeFi Bench website. Here is a sample from the Trader prompt to give a sense of the level of detail each agent receives:

You are the Trader agent, an autonomous DeFi portfolio manager competing in the DeFi Bench yield competition. You control a Makina machine, a USD-denominated (USDC), multi-chain yield strategy deployed on Ethereum mainnet, Base, and Arbitrum. Maximize total portfolio’s share price in USDC by the end of the competition period. You are competing against other frontier AI models on identical machines with the same starting capital and the same approved instruction set.

Every document each agent generates is viewable on the site. Anyone can read what Claude’s Trader thinks about yield sustainability on Morpho, or how Gemini’s Risk Researcher weighs exploit history when sizing an allocation.

A key design decision in Season 1 is the intentional minimalism of the harness. Agents have no access to external tools, MCPs, or web search. Each agent receives one prompt per role per round and must produce a complete structured response in a single pass. This isolates pure model reasoning as the variable under test. When an agent misreads a risk or miscalculates a position size, that failure reflects the model’s own capabilities, not the quality of its tooling.

Why Makina

All execution happens through Makina, the DeFi Execution Engine that already powers Dialectic flagship Machines (DUSD, DETH and DBIT). Makina is vault infrastructure for programmable onchain asset management. It enables operators to deploy, operate and distribute institutional-grade tokenized strategies in a fully non-custodial manner, with strong risk controls and unparalleled efficiency.

Each competing model operates through its own Machine, which is a strategy-specific vault smart contract deployed on Ethereum. The Machine handles deposits, withdrawals, share price calculation, and fee management. Each Machine connects to one or more Calibers, which are the chain-local execution engines where assets are actually deployed.

Calibers contain an instance of MakinaVM, a scope-limited, generalized onchain engine that executes pre-approved smart contract Instructions. The permissioning model uses a Merkle tree of hashed commands and selected parameters stored onchain. To execute any instruction, the Operator (in this case, each agent) must provide the corresponding Merkle proof.

What this means in practice is that even if a model produces an unsafe or malicious command, MakinaVM will not allow it to be executed. The agent has full creative range within the approved universe and zero range outside of it.

Makina is an excellent infrastructure to entrust autonomous agents with capital to manage; it makes it possible to do so safely, and to define a strict environment for the agent to evolve within.

Why Venice

All model inference runs through Venice.AI. Venice provides access to frontier open-source and proprietary models with minimal or model-dependent content filtering, which matters for DeFi Bench because content filters can interfere with financial reasoning. A model that avoids discussing downside scenarios or risk exposure is not useful as a portfolio manager.

Routing all inference through one provider also levels the playing field. Performance differences between models reflect genuine capability rather than variation in hosting infrastructure or API latency.

How a Round Works

Rounds run every eight hours. If it’s an IC review round (every 5th round), the Investment Committee runs first and may update the Trader’s Strategy Document. Then the Risk Researcher produces a fresh risk assessment. Finally, the Trader receives the Risk Note, IC feedback, and a full context package that includes:

• Portfolio state: Current positions with USD values, idle token balances on each chain, and pending rewards, all read directly from onchain contracts.

• Market data: Live APR, TVL, and historical yield trends for every available pool, plus risk metrics covering asset exposure, liquidity profile, and protocol concentration.

• Available actions: The full set of pre-approved Instructions across Morpho, AaveV3, Curve, and Fluid, specifying which tokens, vaults, and actions the agent can execute.

• Performance feedback: Previous round’s PnL, per-position value changes, and the current leaderboard with competitor allocations (but not their reasoning).

Based on this context, the Trader produces a reasoned analysis and an ordered set of commands: supply, withdraw, swap, bridge, harvest. Commands are validated against a root file of approved actions. Invalid commands cannot be executed. Valid commands execute onchain on the Machines’ Calibers..

Agents also maintain persistent feedback loops across rounds. For example:

• The Trader’s Strategy Document evolves with each cycle. Traders can even leave structured self-notes to track cooldown timers or planned multi-step rebalances.

• The Risk Note references previous assessments to track how conditions change.

• IC assessments create a governance layer that feeds back into both the researcher and the trader.

Scoring

Scoring comes from the Machine’s Share Price, a core function of Makina. The Share Price reflects the total value of all assets held by the Machine. Each Machine’s NAV is divided by shares outstanding to produce a per-share price in USDC, pulled directly from onchain accounting. Returns, drawdowns, and Sharpe ratios all derive from the same source. Nothing is self-reported, calculated off-chain or adjusted.

What Comes Next

Season 2:
Agents get access to tools such as web search, external data calls, and selected MCPs. Calibers expand to @base and @arbitrum for multichain execution. The protocol universe grows to include @pendle_fi PTs, LP, and more sophisticated instruments. This will test whether the bottleneck in Season 1 is reasoning or information access.

Season 3:
Agents can propose new yield opportunities for addition to the approved universe. When an opportunity gets approved, it becomes available to everyone, and the agent that found it first gets access first. This rewards genuine research and sourcing ability on top of allocation skills.

We will publish analysis as the seasons unfold. One of the things we are watching closely is whether models converge on similar strategies or genuinely diverge as their Strategy Documents evolve over time. Every piece of data is public, so anyone can form their own view as the competition progresses.

Get Involved

DeFi Bench is live. Leaderboard, analytics, each agent’s strategy documents, the full seed prompts, and the complete protocol universe are all public.

DeFi Bench is an open experiment. We want informed participants following along, questioning the results, and contributing to the conversation about what AI frontier models and agents can actually do in DeFi.

If you have comments, questions, or like what we’re doing, Interact with our posts on X or come talk with us. We usually hang out in the Makina Telegram.

Built by Dialectic, powered by @makinafi and @AskVenice

Recent events show that DeFi will evolve towards an agentic future. Dialectic has been building the defining risk-adjusted yielding machine for 5 years, delivering industry leading risk-adjusted yield since the dawn of DeFi. We were the first to have fully automated execution in 2020, pioneering novel approaches for secure onchain execution, we were the first to have institutional grade risk management in 2022 with onchain insurance, the defining dataset in DeFi for machine learning and analytics as well as preemptive risk protection.

Today we announce that we have achieved fully agentic yielding enhancement, covering the entire process of onchain capital allocation from scouting through due diligence and research, execution, harvest, reporting and audit. Over the last few months our team has developed multiple specialized harnesses for specific tasks, fine-tuned prompts and developed custom MCP tools and skills. We have effectively distilled and condensed over 5 years of DeFi experience, knowledge, data events and operational processes. This fundamentally changes the complexion of onchain asset management and unlocks a new universe of opportunities for crypto-native compounding. We note that Dialectic’s Machine has been a live Machine, producing compelling yields for investors for years. As of today, humans act only in a checks-and-balances capacity, verifying and approving outputs and fine-tuning the agents and their harnesses, with more automation every passing week.

Our focus remains on leadership in onchain risk management as our systems have proven impeccable protective performance over the adverse events of the last several years (10/10 liquidation cascade, stablecoin depegs, various hacks and exploits, etc). This accelerated development has been enabled by the groundwork that our engineering team has laid over the years building the robust infrastructure and back office systems that now allow us to obviate scouting, risk scoring, deployment, harvest fund administration, and audit.

In addition to our core yielding funds, Dialectic will launch many new yield-generating vehicles in the coming months including looped RWA strategies on Makina, Swiss-franc denominated yield products, tranched yielding with Royco, and dozens of other products with strategic partners enabled through our agentic yielding software stack.

As but one example of the Dialectic Machine’s extensible data and AI capabilities, today we announce the DeFi Bench - a real world benchmark test between leading foundation models yielding onchain, as a reference for current model capabilities in the DeFi domain as a real world use case. We feel that this type of practical financially driven race is another interesting way for AI models to distinguish themselves. We note that the DeFi Bench and the Dialectic Machine utilize Makina’s DeFi infrastructure and Venice’s private inference extensively as an indispensable tool for AI using crypto primitives. Makina and Venice are amazing.

Dialectic combines machine intelligence with crypto-native primitives to build superior financial compounding machines. The machines are compounding and this acceleration will remind the world why token-enabled environments are the correct infrastructure for the AI-enabled future of finance.

We are hiring talented people to help us build industry-defining automated compounding machines. If these concepts are interesting to you, please apply on our jobs board.

Dialectic is now the curator for Royco Dawn, Royco’s onchain tranching protocol that splits yield sources into Senior tranches (protected) and Junior tranches (first loss capital). The first vault, Senior Royco USDC (srRoyUSDC), is live and generating yield today, with two more products coming up shortly after: an ETH carry vault (roywstETH) and a leveraged ETH loop vault (sroywstETH). All three vaults are curated by Dialectic and leverage the capabilities of Makina’s DeFi execution engine.

Why Dialectic as the Curator

Dialectic has been running active DeFi strategies since 2020 and is now managing three private funds (Chronograph, Moonphase, Calatrava) across USD, ETH, and BTC with 50+ live positions at any given time. Track record matters because curating the Royco Dawn vaults is not a passive allocation job. The curator runs active strategies across three interconnected vaults, manages collateral positions and leverage loops, makes real-time calls on coverage ratios, and rebalances across all three products.

This type of active vault management needs appropriate infrastructure, risk tooling, and experience navigating the full range of scenarios crypto and DeFi have surfaced over the years.

In October 2025, Dialectic became the first operator on Makina (the DeFi execution engine for onchain strategies), bringing similar strategies onchain for the first time through DUSD,DETH, and DBIT Machines (think vaults with superpowers). These new strategies are available for anyone to interact with onchain, and demonstrate Dialectic’s ability to operate public vaults while maintaining their own high standards of operations.

Operating Royco Dawn vaults is a natural extension of their combined experience of managing private strategies internally and public vaults via Makina. Now, as curator of the Royco Dawn vaults, Dialectic is bringing its expertise, risk management practices, and Makina infrastructure to a structured product featuring tranching, collateral management, and leveraged loops.

How Royco Dawn Works

Royco Dawn takes yield sources like lending markets, staking protocols, and tokenized RWAs and splits each one into two risk tranches: a Senior tranche and a Junior tranche. Each of these tranched yield sources is called a market.

Both tranches deploy capital into the same underlying strategy at the same time. Junior capital is co-invested alongside Senior, earning the same base yield from the same source. The difference is what happens if something goes wrong. If the underlying position draws down, experiences a smart contract exploit, an oracle failure, or a liquidation shortfall, Junior absorbs the loss first. Senior depositors aren’t affected unless the Junior’s entire balance has been brought to zero.

The size of that protection is defined by a coverage ratio and enforced onchain. A 20% coverage ratio, for example, means the Junior can absorb up to a 20% drawdown on the total pool before the Senior is exposed. If coverage drops below the minimum, new Senior deposits are paused until Junior capital is restored.

The yield split between the two tranches isn’t fixed. It adjusts based on how much Junior capital is available relative to Senior deposits. When the Junior is scarce, the protocol pushes more yield toward the Junior tranche to attract first-loss participants. When there’s plenty of Junior capital, more yield flows to Senior. The split also self-adapts over longer timeframes. If Junior stays scarce for days or weeks, the premium keeps climbing until the economics are attractive enough to bring new capital in.

As curator of the Royco Dawn vaults, Dialectic selects which markets to allocate into, which tranches to take, what coverage ratios to run, and how to rebalance as conditions change. Depositors can get exposure to the strategy without having to evaluate every market themselves. In the case of the first vault, srRoyUSDC, Dialectic only allocates into Senior tranches across the markets it deems worthy.

To go deeper into the Royco Dawn mechanics, tranching, and risk frameworks, refer to Royco’s docs.

The Three Vault Products

Dawn Senior Vault - srRoyUSDC: is the first vault and is now live. It gives USDC depositors diversified yield exposure with built-in first-loss protection on every position. Dialectic allocates that capital across a set of Senior tranches from Royco’s market menu (see pic below), where each underlying position has its own Junior buffer absorbing losses before Senior is affected.

ETH Carry Vault - roywstETH: is coming next. It lets wstETH holders earn yield on top of their staking rewards without selling their ETH. Dialectic takes the deposited wstETH, posts it as collateral on lending venues, borrows USDC against it, and deposits that USDC into the srRoyUSDC vault.

ETH Loop Vault - sroywstETH: is the leveraged version and will arrive after roywstETH. It’s for depositors who want amplified exposure to roywstETH’s yield. It deposits wstETH into the roywstETH vault, receives roywstETH tokens, posts those as collateral, borrows ETH, converts it back to wstETH, and repeats. This loop aims to maintain a 2-4x leveraged exposure to the carry vault’s yield.

The three vaults form a layered structure with srRoyUSDC as the base layer generating Senior tranche yield. roywSTETH builds on top of it, and sroywstETH adds leverage on top of roywstETH.

For further reading on the vaults and the associated risk please refer to the Royco Docs.

Vault Management Under The Hood: Leveraging Makina

Dialectic is using Makina for the infrastructure and operations of the Royco Dawn vaults. Makina lets professional managers run onchain strategies through non-custodial vaults called Machines, each with a defined set of pre-approved actions (called Instructions) that constrain what the curator can do with the capital inside it. Dialectic has been running strategies on Makina since October 2025 as the protocol’s first operator, managing DUSD, DETH, and DBIT. The Royco Dawn vaults will be managed using the same infrastructure.

For vaults running strategies on Royco Dawn, the pre-approved Instruction set is a natural fit. Dialectic’s mandate as curator is specific. Operate within the Dawn markets, allocate into Senior tranches, manage coverage, and rebalance. The Instruction set for the Royco Machines reflects exactly that scope. Dialectic can execute everything the curator role requires, and nothing beyond it.

If the Instruction set ever needs to be updated, the change goes through a timelock where it can be reviewed and challenged before taking effect. This means if someone tries to push through an action that falls outside the Dawn market scope, it gets flagged before it goes live. For depositors, that means hard boundaries of what Dialectic can do with the capital, enforced onchain and verifiable.

The pre-approved nature of the Instruction set also enables Dialectic to automate operations. Because every allowed action is already defined, Dialectic can operationally handle rebalancing, allocation updates, and position maintenance programmatically through Makina’s powerful CLI tool. The alternative would be routing every operation through a multisig or MPC wallet, collecting signatures each time a position needs to move. For infrequent transactions this may work, but these vaults require active management across multiple markets with leverage loops and collateral positions that need constant attention. Operating the vaults through Makina’s CLI means Dialectic is much more efficient and can keep up with strategy demands. For depositors this translates into tighter vault management, stronger optimization and ultimately better returns.

The Royco Dawn vaults benefit from the same infrastructure Dialectic uses across all of its onchain strategies. Non-custodial, cross-chain execution through Makina’s hub and spoke architecture, and a coherent operational layer that empowers the manager to do more. Combined with an Instruction set that constrains execution to the curator mandate and execution that keeps pace with active strategy demands, Makina provides Dialectic with the appropriate infrastructure to operate the Royco Dawn vaults at the standard they’ve set across other strategies.

Get Started

The first vault, srRoyUSDC, is now live and accepting deposits. The roywstETH and sroywstETH will follow shortly.

For more on the tranching mechanics, vault architecture, and risk framework, see the Royco Dawn docs: https://royco.gitbook.io/royco-dawn

For more on Dialectic’s thinking about the crypto space, fund information, and approach to onchain strategy, visit Dialectic.com

For more on Makina and what makes it the most advanced vault infrastructure in DeFi, visit makina.finance

Original Publication by: Felix_v_H • January 9, 2026

Introduction

The journey from private crypto investments to accessible, institutional-grade DeFi infrastructure has been long and fragmented. For years, sophisticated yield strategies were only available to accredited investors and had high minimum liquidity requirements. Retail participants settled for basic lending yields of 3-4% on platforms like Aave or much riskier and complex opportunities. Until now, the lack of high quality and accessible onchain strategies has limited DeFi’s long-term growth.

As the first Operator on Makina, Dialectic has witnessed firsthand how this new execution engine bridges that divide. Makina represents the infrastructure layer that will define the next era of professional DeFi asset management. Since the prelaunch phase in late September, we’ve seen the market validate what we believed from day one: there is massive demand for transparent, professionally-managed onchain strategies that combine institutional rigor with DeFi accessibility. Makina combines sustainable economics, network effects, great strategic positioning, proven market demand, all with a focus on security and transparency.

The Evolution of DeFi Vaults

DeFi is historically plagued by liquidity fragmentation across chains, protocols, liquidity pools etc. For sophisticated players, these inefficiencies create great investment opportunities, due to advantages in opportunity scouting, due diligence and live risk monitoring. Vaults offer a convenient middle layer for end users and institutions alike, enabling access to sophisticated yield strategies without the complexity of manual deployments.

Despite recent volatility and events like the Stream Finance collapse that sparked conversations about transparency, these discussions fail to address fundamental structural limitations. Existing vault products are either passive aggregators held back by lengthy governance processes, lack cross-chain capabilities and require extensive manual integration work.

BoringVaults and ERC-4626 that comprise most vault TVL today (figure 1) and incoming ERC-7540 based vaults all address specific pain points but fall short in combining the professional Operator model, cross-chain execution capabilities, institutional risk controls and transparency that the market demands.

The Blue Chip Thesis

Attaining blue chip status in DeFi requires more than temporary success. The vault infrastructure vertical remains nascent—no project has achieved blue chip status yet. Standing amongst industry giants like Aave, MakerDAO or Uniswap requires sustainable economics, network effects, strategic positioning and proven security. We believe Makina is exceptionally well positioned to fulfill all of these demands.

Why Makina’s Architecture Matters

Makina’s fundamentally different architecture represents a new approach for vault operations. At its core lies the hub-and-spoke model of a Machine deployed on Ethereum Mainnet as the hub, interacting with Calibers serving as execution environments on Layer 2 chains (figure 2). This design enables transparent unified strategy management across multiple networks, quick deployments on new chains, all without sacrificing security benefits of non-custodial operations.

MakinaVM, the execution environment for strategies, eliminates the need to write custom smart contract adapters for each protocol integration. As cyber•Fund noted in their Makina investment thesis, this approach means “every protocol integration no longer requires new code, new audits, nor does it introduce new attack vectors.” Any blueprint to interact with smart contracts that has been created by an Operator is saved in a public repository and usable by all other Operators.

This architectural flexibility extends to risk management. Unlike existing vaults where risk controls are either non-existent or too rigid, Makina enforces onchain risk parameters such as loss thresholds and cooldowns for various operation types, governed by the risk manager. The risk manager pre-approves transaction types, ensuring protocol integrations align with the strategy’s mandate. This prevents Operators from taking excessive risks while giving them maximum flexibility to manage strategies effectively.

Sustainable Economics

Makina’s Operator approach enables professional asset managers to deploy their expertise fully onchain. Value created is distributed among all participants: liquidity providers receive the majority of returns for providing capital, Operators charge management and performance fees aligned with their strategy type, and the Makina DAO receives a share of those fees that accrue to MAK token holders. Strong performance attracts new capital and benefits all involved parties.

Network Effects

As Makina grows and proves its architectural effectiveness, more professional Operators will launch their own Machines. Each additional Operator brings expertise, reputation and capital relationships, increasing Makina’s attractiveness for users seeking diverse strategy options tailored to their risk appetite. Machines can be integrated into other applications to expand reach beyond native DeFi users, including (crypto) Neobank apps. In that sense, we believe that the total addressable market (TAM) of Makina is equal to or at least highly correlated with the TAM of onchain asset management. As crypto and traditional finance converge, Makina is poised to be the best execution layer for the future of asset management.

As standard ERC-20 tokens, Machine Tokens are fully composable in DeFi. Dialectic’s Machine Tokens DUSD, DETH and DBIT are already integrated on Curve, Pendle, Morpho and Gearbox, with more integrations coming. Operators can freely decide which integrations best align with their strategy. Makina will grow alongside DeFi as the opportunity set expands thanks to the chain-agnostic protocol design of MakinaVM.

Strategic Positioning

Makina occupies the intersection of multiple power trends: institutional DeFi adoption, professional onchain asset management, RWA tokenization, cross-chain execution and transparent onchain operations. This provides multiple paths to sustainable growth and resilience against competitive threats.

As Operators ourselves, we avoid betting on which chain will dominate and prefer deployment flexibility. Chain TVL fluctuates drastically over time, with new incumbents coming to market regularly (figure 4). Native cross-chain execution provides diversification benefits and ensures Makina is positioned for any eventuality in protocol and chain dominance over time, always positioned to where the best incentives are.

Market Demand

Fueled by the pre-launch campaign, Makina saw rapid TVL growth, surging to $94m in the first week. The protocol showed resilience during market turbulence in late October that saw the broader sector’s TVL drop by almost 50%, despite the initial incentive campaign ending around the same time.

Focus on Security

Security itself will be proven over time as Makina achieves more lindy. However, managing assets onchain requires more than bulletproof smart contracts. It extends to the entire package given to Operators and end users: clear investment mandates and adherence to such mandates, onchain accounting, the security module, recovery mode, the security council and atomic execution. We refer to the Makina Docs for further details on any of these aspects.

Conclusion

We believe vaults will be one of the main entry points for users into DeFi, offering use cases ranging from simple savings accounts or high-risk actively managed strategies deployed according to individual risk appetites. Within this context, Makina solves fundamental problems to support the growth of underlying infrastructure to enable such a future.

The case for Makina to evolve into a blue chip protocol rests on five pillars outlined above. From Dialectic’s perspective, the question is not whether we should continue growing on Makina, it is how quickly we can scale our existing strategies and create new ones to serve the demand that we know exists. The technology is ready, the capital is waiting and the market is validating the approach. The journey to blue chip status has begun.

Original Publication by: Ogst • July 15, 2025

Yield farming, a core mechanism within Decentralized Finance (DeFi), has rapidly evolved alongside the broader expansion of blockchain-based financial systems. Once a niche activity, it now plays a central role in capital allocation across DeFi protocols. However, the complexity and constant innovation in this space make it challenging to develop a clear understanding of its mechanics and associated risks.

In this Systematization of Knowledge (SoK), we present a structured overview of the yield farming ecosystem and its vocabulary through three core dimensions: yield sources, yield quality, and associated risks. This document aims to map the current DeFi landscape in which Dialectic operates, offering clarity and perspective on a rapidly evolving domain. Rather than delving deeply into each topic, our objective is to articulate and formalize the vocabulary we use internally, establishing a shared conceptual framework for navigating the space. We intend to curate and maintain this post as the DeFi landscape evolves, ensuring it remains a relevant and living reference.

Yield sources

Let us begin by outlining the primary sources of yield generated through DeFi activities. These represent the core strategies currently composing the Dialectic yielding books.

- AMM (Automated Market Maker): Facilitates decentralized trading by allowing users to provide full-range liquidity to token pairs, earning trading fees and, in some cases, protocol incentives. Examples: Uniswap v2, Curve.

- clAMM: Enables liquidity providers to allocate capital within a specific price range, significantly increasing capital efficiency. While returns are structurally similar to those of simple AMMs, primarily derived from trading fees, they require active management to optimize performance and mitigate impermanent loss. Examples: Uniswap v3, Aerodrome, Velodrome.

- Basis Trade: A delta-neutral strategy that captures the funding rate differential by shorting a perpetual futures contract while simultaneously taking a long position in the corresponding spot asset for the same notional amount. This structure allows for yield generation with minimal directional exposure. Examples: Ethena, or manually executed strategies.

- Bridge Passive Liquidity: Provides a buffer of liquidity to facilitate cross chain transfers in exchange for fees and incentives. Examples: Across, Hop, Connext V1.

- Lending: Involves overcollateralized on-chain borrowing and lending, where users supply assets to earn interest while ensuring solvency through excess collateral. Examples: Aave, Morpho, Euler.

- NFT Lending: Facilitates lending markets where loans are overcollateralized by non-fungible tokens (NFTs), typically high-value or blue-chip assets. This emerging category introduces unique valuation and liquidation challenges. Example: Gondi.

- Looped Lending: A leveraged yield strategy that recursively supplies and borrows similar assets to amplify returns, profiting from the spread between supply and borrow rates.

- PT – Fixed Rate (Principal Token – Fixed Rate): Fixed-rate yield products structured by splitting the yield-bearing asset into principal and yield components. Users lock in a fixed return by purchasing the principal token, with the underlying yield sourced from existing protocols. Examples: Pendle, Spectra.

- Savings Rate: Low-risk or near risk-free yield derived from staking assets in protocols that passively distribute rewards, often backed by conservative underlying strategies such as tokenized treasury bills. Designed for capital preservation with minimal active management. Example: sDAI.

- Junior Tranche: The subordinate portion of a structured product that absorbs losses before senior tranches, in exchange for higher potential returns. Typically suited for risk-tolerant participants seeking enhanced yield. Example: Aave Umbrella.

- Perps Passive Liquidity: Capital is passively deployed to provide liquidity on perpetual futures exchanges, effectively taking the opposing side of traders’ aggregate PnL. This strategy can generate returns over time, particularly during periods of elevated volatility and frequent liquidations. Examples: Synthetix V3, Umami, GMX.

- Options Selling: A short-volatility strategy that involves selling call or put options in exchange for a premium. Examples: Cega, Ribbon.

- Crab Strategy: A delta-neutral, short-volatility options strategy designed to generate yield when an asset’s price stays within a defined range. Returns are maximized in low-volatility environments, while sharp directional moves can lead to losses. Example: Opyn.

Associated risks

Participating in DeFi comes with its own unique set of risks. Understanding these is essential for evaluating strategies and managing exposure effectively:

- Directional Exposure: The risk associated with holding unhedged or biased positions that are sensitive to market movements. Such exposure can lead to significant losses during adverse price swings.

- Smart Contracts: Faulty or vulnerable smart contracts can be exploited by malicious actors, exposing funds to risk in DeFi.

- Oracle Manipulation: Oracles are integral to many DeFi protocols. When a price oracle is manipulated, attackers can extract value from users. Example: Manipulating the USDC/WETH price oracle on Aave could result in fraudulent liquidations.

- Governance (Malicious or Compromised): Protocol governance entities that control smart contract logic may act maliciously or be compromised. This can lead to severe exploits. Example: The Radian exploit.

- Bridges: Liquidity on Layer 2s or other chains is often not native and relies on bridges for backing on the mainnet. If the bridge is attacked, assets may be at risk. Example: The Nomad Bridge hack.

- Operational Risk: Since DeFi is based on self-custody, poor wallet management or security (e.g., phishing or private key leaks) can lead to loss of funds. Example: Bybit hack.

Yield quality

Beyond capital risk, it is essential to assess the quality of yield. This includes evaluating the reliability, consistency, and sustainability of realizing projected returns.

- Paid in Kind: Returns are paid in the same numeraire as the allocated capital. Example: USDC lending on Aave where yield is paid in kind on a block-by-block basis.

- Liquid Tokens: Incentives are paid in existing liquid tokens. Example: Aerodrome DEX pays liquidity providers with the AERO token.

- Locked Tokens: Incentives are paid in tokens that are vested. Example: Euler pays with rEUL.

- Points: Returns are paid via points accumulation, representing a future amount of tokens. Points are the hardest category to assess, as they are challenging to price in terms of the underlying numeraire.

Original Publication by Ryan Zurrer • April 28, 2025

As we celebrate our 5th anniversary, I thought it pertinent to consolidate our views on what makes working at Dialectic special. I often describe Dialectic as a compounding machine. But it isn’t just an on-chain machine for compounding capital. The core mechanisms in place are designed to allow great people to compound on their personal development and entrepreneurial skills, seek truth together, and build the Dialectic compounding machine as a cohesive, high-performance team.

At this stage of our evolution, I observe 6 primary categories that make working at Dialectic quite unique:

1. We are crypto-native entrepreneurs

Dialectic was built by crypto-natives for crypto-natives. Our core team members were the first primary investors in our suite of funds and the industry-leading products and reporting that we’ve built were largely born out of a desire and need to have superior risk-adjusted compounding on our own crypto assets. We built Dialectic for ourselves and discovered along the way that there are other crypto-natives who value what we do. Everyone at the company approaches their work with an entrepreneurial mindset and takes on a significant level of ownership and responsibility. Everyone participates in developing our strategy as we genuinely take a dialectic process to arriving at truth, getting in sync, and orienting our compass for the journey ahead.

2. The Dialectic Process

Intellectual curiosity, experimentation, and debate in the pursuit of truth is foundational to building the inspiring future we want to live in. Embedded in our daily work is what we refer to as the Dialectic Process, a commitment to objectively seek truth in this rapidly evolving landscape. We achieve this through thoughtful debate and open-minded intellectual experimentation. I am consistently impressed by how our participatory approach to strategy, whether at offsites or during regular sync meetings, lead us to arrive at nuanced, often previously unexpected realizations. I frequently walk into a meeting with one assumed perspective and leave aligned with the team around something richer and more evolved, shaped by the collective insights and diversity of thought in our group. There is a deep amount of mutual trust and understanding that has been built up by compounding on credibility and the relationships in this group over time.

3. The Dialectic Programme

Deeply connected to our culture is a structured programme of self-reflection and commitment to personal development, both from the perspective of mental health and self-awareness, but also skills and abilities which then feed into the machine that is Dialectic. Twice per year we conduct offsites that start with lessons learned sessions over the previous semester and culminate in strategies and goals at the corporate level distilled down to each individual. Annually, we all individually compose a deep dive personal journal that covers all corners of one’s life. This intense journaling process has driven incredible growth in our young staff members and as a result we take these reflective activities seriously and give time off to staff to allow for focus on this deep personal work. Further, everyone undertakes a personal development journey parallel to their work. That may be exploring mental health strategies, learning novel AI tools, emerging adjacent fields such as quantum or new technical skills. Each employee chooses their own adventure for each year. This is paired with regular iterative coaching and feedback sessions with peers which allows the team to learn from each other. Overall, the Dialectic Programme is designed to allow great people to compound on lessons learned and intellectual curiosity.

4. The Dialectic Compensation Framework

An elegant approach to sharing value with those who create value:

I often say ‘it is way more fun to get rich with your friends’. I have designed a compensation framework that has 4 pillars and allows employees a path to become meaningful owners of the business and build life-changing wealth along the way. We work really hard and building durable value is difficult. So it is important to me that there is a clear path for great contributors to build ownership in what they’re working so hard on. The 4 pillars of compensation are:

1. Salary - our baseline compensation alone is competitive with the industry and takes into account the cost of living in wherever team members are based

2. Discretionary annual bonus - we share a significant component of annual cash flows based on how the company has performed and how individuals have contributed to that performance in the given year. Annual bonus is composed and reviewed by the board and senior staff for fairness and uses objective metrics

3. Equity - people who are successful at Dialectic build ownership over time based on a specific framework in which our team is committed to and very happy with. I don’t want employees. I want committed partners and entrepreneurs around me because that makes business building more fun and accelerates growth for all of us

4. Entrepreneurship & Intrapreneurship - at times employees will want to start something new, either within Dialectic or beyond. Since it is important to me to walk the talk of being a platform for entrepreneurship, we take an open-minded view towards experimentation in this regard and often are the first investor and collaborator when employees transparently present new and exciting projects. 1OF1, our industry-leading art collective and Makina are good examples of this fourth pillar.

5. The Dialectic Culture

My favourite characteristic about Dialectic is that we have a remarkably talented team of people that bring a lot of fun to their work, derive meaning from what we’re building and care for each other.

We work with the best. Our group is composed of wicked-smart professionals, who are intellectually curious about the world, open-minded builders in crypto, future-forward and ultimately self-reflective. Humour is a core part of our values and every day I find myself laughing out loud. Inside Dialectic the learning curve is really steep and we work really hard, but what we do is really fun and everyone is passionate about it. Who wouldn’t be passionate about building a machine that literally makes money? We also derive great meaning from the impact that we have. The Dialectic Reciprocity Pledge is a good example of this. Finally, we take care of our group – Dialectic supports personal fitness training for every employee, we provide an amazing set of travel experiences that we share together and enjoy the relationships that we have with hilarious team dinners and regular social events. We remember to mix in epicurean delights with the stoic path of building and entrepreneurship. ‘Meaningful work and meaningful relationships’ is a Dalio principle that we apply within our organization.

6. Compounding is at the core of everything we do.

Our founding thesis is that compound returns are the strongest force in the universe. We know this is true in nature with the golden ratio, we know it is true in finance with compound interest. But Dialectic has also shown that this is also true in relationships, skill development and code. Our core founding team remains intact, developing at a remarkable rate both personally and professionally and the trust we have in each other has grown incredibly strong as a result. We have also observed that our codebase continues to build on itself and all of this has led to the best yielding shop in crypto.

If this sounds awesome to you and you think you have what it takes, please apply via our jobs board or reach out to our team to get involved.

Ryan Zurrer,

Founder

Original Publication by: Ogst • October 1, 2024

Nebula is a graph-based representation of DeFi, designed to model the composability nature of decentralised finance for risk assessment and navigation purposes.

In this blogpost we propose to leverage the developed tool to decompose and analyse the yielding portfolio of a popular DeFi user, namely Bizyugo.

Note that given our domain of expertise, we exclusively focus on BTC, Stablecoins and ETH denominated pools. Other assets and protocols are discarded from the analysis.

Introduction

Bizyugo is a prominent DeFi farmer, managing a wallet that ranks among the top in the space. Currently, it holds 5th place on DeBank and 3rd on DegenScore. This wallet has been active for an extended period, with a diverse range of positions spanning across the DeFi ecosystem. Its portfolio is particularly notable for its breadth, consisting of 78 protocols, 12 chains, and 135 assets. In this post, we use our internal tool, Nebula, to dive into Bizyugo’s portfolio and uncover potential inefficiencies.

First-Order Dependencies: The Tip of the DeFi Iceberg

When evaluating a DeFi portfolio, the first step typically involves loading it into a preferred portfolio management system (PMS), such as DeBank, where high-level metrics such as number of protocols, chains, assets and AUM are readily visible. However, this perspective does not offer visibility into the portfolio’s more complex, n-order exposures.

For instance, the manager might fail to recognize exposures to a specific oracle they find unreliable. Furthermore, as the portfolio expands, the likelihood of such oversights increases, potentially resulting in poorly informed allocation decisions.

Nebula (reminder & updates)

Nebula is introduced and detailed in Introducing Nebula : A graph based representation of DeFi as a knowledge graph representing DeFi. Since the last blog post, we have updated its structure. The major additions being:

• Added Contract nodes representing Asset and Pool implementations, to which they are connected through the IMPLEMENTED_AT relationship.

• Added Wallet nodes representing addresses that are not contracts themselves.

• Added (Pool)-[:CONTAINS]->(Pool) relations, utilized to represent the relationship between gauges and liquidity pools, as seen on Curve Finance, where LP tokens can be staked in a gauge contract or a Convex pool.

• Added CEX nodes representing centralized exchanges, backing tokens like cbBTC, cbETH, BBTC ...

The current structure of the graph is defined by the following schema :

As for Bridges, one could argue that pool nodes should be directly connected to the Oracle they depend on. This functionality will be included in the upcoming Nebula update.

Going forward, the color coding for nodes is defined as follows:

Bird’s Eye View

We integrate the subject’s wallet into the Nebula graph as a node, linking it to each of its positions (Pool nodes) via an ‘IS_IN‘ relationship. Leveraging the graph’s pool-centric architecture, we can efficiently map all the wallet’s dependencies by executing the following Cypher query.

The query identifies the newly introduced wallet node (”Bizyugo”) and initiates a Breadth-First Search (BFS), traversing all possible directed edges originating from the wallet. This process maps the entire network of dependencies, providing a comprehensive view of how the wallet is interconnected with other entities within the graph.

N-order dependencies

To address the blind spots of traditional PMS, we leverage Nebula across four different risk assessment scenarios to reveal hidden exposures.

To quantitatively propagate the exposures through the graph, we define the following procedure:

1) Set the usd_exposure field to 0 for all nodes in the graph.

2) Set the usd_exposure field for each pool connected to the wallet, with the corresponding dollar value allocated to it in the portfolio.

3) For each pool with usd_exposure > 0, increment the usd_exposure of all nodes connected (through an arbitrary number of directed edges) to the pool by the pool’s usd_exposure amount.

In terms of Cypher queries, it translates to the following set of instructions:

1. Assets exposure

N-order asset exposure refers to the indirect exposure a portfolio has to a crypto asset through another DeFi entity (Protocol, Asset, Oracle...). For instance, if stablecoin A is backed by asset B, there is an inherent dependency on asset B. Depending on the nature of asset B backing A, the portfolio manager might not be comfortable entering the position, as the security profile (including risks related to the smart contract, governance, and economics) of asset B also impacts asset A.

In the following example, we break down an allocation of deUSD to an Abracadabra pool, which introduces additional dependencies on Maker and Lido through deUSD’s collateralization mechanism. If Maker fails, deUSD will be affected as well.

Even though Lido and Maker are considered trusted protocols in the ecosystem, the accumulation of these N-order asset exposures within a portfolio can distort the manager’s perspective, leading to an underestimation of the exposure to various entities and the correlated risks across positions.

2. EOAs exposure

Another aspect of DeFi closely monitored at Dialectic is the governance risk of the book, coming from EOA (Externally Owned Account) exposures. Key smart contracts owned by simple wallets represent major points of failure in the space. Instead, we place more trust in ownership through reasonable timelock contracts and multisigs, which we can actively monitor.

In the query below, we reveal the subject portfolio’s exposure to one EOA address through the SOLV protocol. Using the capital propagation method described earlier, we calculate a total of $120,000 worth of BTC (2) that could potentially be siphoned by this wallet, resulting in a 0.4% loss at the portfolio level, which currently represents the equivalent of one month of yield in the BTC-yielding ecosystem.

The (Solv protocol)-[:IMPLEMENTED_AT]->(contract) relationship represents an implementation contract of the Solv protocol, which could be a router, factory, oracle, or any other contract used to implement the protocol logic.

3. Oracles exposure

Oracles are a crucial part of DeFi, controlling lending markets and other protocols. When they lack sufficient resilience, they can be vulnerable to manipulation, causing significant losses for unsuspecting liquidity providers. Some oracles are more reliable than others, with Chainlink being our preferred choice over custom implementations. Nebula enables us to measure and track exposure to this risk.

Below is an example of the top3 dependency on the Chainlink oracle obtained through AAVE, mkUSD and DOLA positions.

4.CEX exposure

Similar to our approach with EOAs, we closely track dependencies on centralized entities like exchanges. The collapse of FTX and others has led to substantial losses and a shaken community, so we are focused on minimizing our exposure to these risks. Nebula plays a key role in helping us identify and monitor these vulnerabilities effectively.

In the following query, we illustrate how one can retrieve exposures of the subject wallet to centralized exchanges:

Risk vs Return

Our main objective being to construct a high-performance portfolio with a strong focus on risk-adjusted returns. Let us leverage our risk scoring algorithm, which uses Nebula as a backbone for risk propagation, along with our returns tracking engine to identify inefficiencies in the portfolio, such as positions with poor risk-adjusted returns.

Our algorithm identified a $5,000 allocation to the DOLA/mkUSD BalancerV2 liquidity pool as the most inefficient position in Bizyugo’s book, with a high-risk score of 60% (see annex [1]) compared to an average return of 1% (see annex [2]). Given both the size of the allocation and the pool’s TVL, which also stands at $5,000, this appears to be a neglected position.

This is a typical position we would rotate into safer or higher-yield opportunities.

Conclusion

We found that the flat view provided by traditional PMS is inadequate for effectively managing large portfolios engaged in DeFi activities. A flat view is inherently skewed and fails to represent the full spectrum of risk exposures.

In this blog post, we utilized Nebula to address this limitation through concrete examples. While our approach was relatively manual, involving the design of queries and analysis of visual results, the true potential of Nebula emerges when external algorithms are integrated to interact with it.

Nebula’s purpose is not to rely on visual outcomes but to enable other algorithms to interact with the graph, producing statistics and guiding investment decisions. From risk management to automated rebalancing, Nebula serves as a data backbone, essential for other programs to safely and efficiently navigate the complexities of DeFi.

Future work

Risk scoring being a central topic in the space of DeFi portfolio optimization, we propose to outline our risk scoring algorithm based on Nebula in the next blog post.

Annex

[1] Text-based representation of the DOLA/mkUSD BalancerV2 pool dependency graph, optimized for easier manipulation within our systems. The base risk scores aggregate to 60% at the pool level, classifying it as High risk (with classifications being Low, Medium, High, or Degen).

[2] Historical APR of the DOLA/mkUSD BalancerV2 pool. We observe occasional spikes in the APR, likely resulting from sporadic trades routed through the pool. A negative APR accounts for impermanent loss or tokens depeg.

Context

As we forge forward into the 4th major cycle in the history of crypto, we’ll again see an influx of TradFi folks taking the trip down the crypto rabbit hole and trying to apply their TradFi frameworks to crypto. In this post and related ones to follow, I offer some lessons learned and considerations on the difference in the required paradigms between crypto and TradFi and how the unique heuristics required in crypto can unlock extraordinary value, but pose a different risk profile to investors.

Crypto-Native Investment Classes

There are four relevant investment categories in crypto assets:

1- Crypto Venture is characterized by early-stage pre-launch projects, usually with some combination of equity and tokens depending on the network and resultant value accrual mechanism. This includes both liquid and illiquid venture. Note: “liquid venture” is an opportunity whereby the given project’s token is live and openly tradable, but the project is still in its infancy and has the characteristics of an early-stage venture project. Liquid venture is truly unique to crypto (outside of the wild west of the Canadian TSX ventures exchange)

2- Late-Stage Liquid Tokens are mature tokens of fully formed communities with an operational peer-to-peer network. This category includes staking opportunities and single-sided yielding. These projects usually have some form of product-market-fit in their offering and offer arbitrage traders, fundamental investors and market-makers different opportunities. Examples are Bitcoin (digital gold) Ethereum (decentralized compute), or Filecoin (decentralized file-storage).

3- Fixed-Income Yield Farming is whereby fixed-income returns are driven by providing liquidity to on-chain capital markets and includes keepers, market-neutral yielding on stablecoins as well as risk-on yielding with risk-on pairs. Examples are Curve/Convex or Uniswap V3 pools.

4- Late-Stage Private Equity (and Public Equities) are mature businesses that have established product-market-fit, and generally profitability and/or sustainability. This is the least interesting category in crypto from a return-profile perspective, but are important organizations in the space. Examples are COIN (public) or Kraken (private, for now, expected listing in the next year). The reason why this category is less compelling is that it is almost always easier to go from 0 to 1 than it is to go from 1 to 10 or 10 to 100. The larger a project is, the more difficult it becomes to deliver a multiple on capital. Given the overall risk profile of crypto, the returns should be expressive imho.

Do note that there is crossover between these categories and they should be thought of somewhat as intersecting venn diagrams. Below I outline a generalized mental map of return profile vs liquidity profile of the different categories.

I highlight crypto venture in this discussion as the delta between crypto venture and traditional venture are most extreme as many of the long-held heuristics of traditional venture get blown up with crypto. Some primary examples are below:

Timeline to Liquidity

Certainly, the most remarkable advantage that early-stage crypto venture has over traditional venture is timeline to liquidity. In a traditional venture deal, if everything goes well, it will take 7-10 years to harvest via a complicated IPO process after many rounds of dilution. Conversely, crypto projects typically go from ‘seed stage” to effectively ‘public’ within months and not years. This allows investors to take a better risk-adjusted approach to early-stage investing by being able to pair back exposure as projects go live. The feedback loop driven from the market allows teams to iterate and evolve much more quickly. However, crypto investors then are faced with a very different harvesting decision process.

Dilution

I am frequently asked why early-stage token-enabled projects command such a significant valuation premium to traditional venture-backed companies. Why are seed valuations well into eight-figure and even nine-figure territory and how does that make sense? It is important to note that a token-enabled network doesn’t have dilution outside of the inflation on the network post-launch. So each successive private round only dilutes the community and team and not previous investors. Due to superior dilution protection, it is natural that valuations be higher. Note a community could come together to change the inflation schedule, but this is exceedingly rare in the space and usually value destructive to a project due to the controversy and broken trust inherent in such a decision.

Value of Community

Crypto projects are uniquely enabled to tap a large community to support early-stage growth and development. Often early-stage rounds in crypto look more like “party rounds” whereby dozens of investors take a position instead of a defined lead with few participants. This is a superior capital model because A) a project can’t predict who will be their best value-add partner at the outset; and B) When an entrepreneur raises from a crowd instead of a select group of VCs, they get a massive amount of free work, product evangelism and support that they wouldn’t get from even the most well-intentioned institutional investor. In the 2020s communities will be some of the most valuable assets on the planet and crypto uniquely unlocks highly incentivized communities who passionately support their project. Conversely, in traditional venture, often the best a start-up can hope for from their community is that it’s a community of customers who are willing to do a single activity such as provide feedback

Liquidity & Risk Profile - Harvesting Strategy

In traditional Venture, the moment of harvest is usually decided for the investors - it’s the acquisition or the IPO. However, with crypto, investors have to ask themselves tough questions about harvest strategy every single day once the token is live, typically early in the project’s lifecycle.

Dialectic conducted a study looking at dozens of SAFTs and their resultant liquidity. We discovered that, overall, the superior harvesting strategy is to gradually sell the SAFT along the vesting schedule. The reason for this is because crypto goes through highly volatile cycles and typically one has future opportunities to invest in a market bottom sometime after a project has launched its open network. This harvesting strategy flies in the face of traditional venture heuristics of “double down on your winners” and power law. For example, almost every L1 and L2 since the dawn of Ethereum has bled value against Ethereum after a 90-180 day “honeymoon period”.

Investors may perform fine vs. standard benchmarks, but if a fund can’t outperform some ratio of ETH and BTC, then they are likely not going to have longevity. Therefore, crypto investors are presented with a more dynamic dilemma of when to harvest - both a compelling opportunity for the prudent data-driven capital allocator and a risk for the emotionally attached early-stage backer.

Portfolio Construction & Deal Sizing

The timeless adage on Sand Hill Road is “concentration matters and double-down on your winners”. This has been true for most of Venture’s history. However, not with crypto. Since the delta in valuation is so expressive in such a short period of time, having exposure to experimental fields is more important than having significant concentration. It is exceedingly difficult to predict which project will win in a given category. A bet that is 1% of a portfolio will almost never be a fund-maker in traditional venture, but often is in crypto. Further, because the public market is an incredibly efficient market and there usually aren’t more than 2 rounds before public launch, more often than not, follow-on rounds are regularly quite dilutive and can be quite destructive to fund returns overall. Thus, diversification matters more in crypto than in TradFi and the results of diversified portfolios often outperform vs TradFi peers.

Hit Rates

Given the aforementioned advantages, the hit rate, defined as the rate of highly successful outcomes over the total number of invested projects for a given investor, is considerably higher in crypto. According to Perplexity, in traditional Venture, the percentage of investments that return positive capital is around 25%, with about 5% returning 10X or more and .04% returning 50X or more. In crypto, the percentage of projects returning 10X or more is approximately 8% and nearly .25% of crypto projects return 50X or more. The quicker market feedback loop through improved liquidity profile are primary reasons for this marked increase. There has never been an industry on planet earth as profitable as programmable money.

Implications & Predictions

We clearly see the advantages of early-stage crypto investing. The increased liquidity of early-stage investments, the faster feedback loop, the ability to recruit a community towards a cause are all reasons why eventually there will be a merging between crypto venture and traditional venture. In the not too distant future, we will see the renaissance of decentralized capital formation or ICO 2.0. With some reasonable legal frameworks in place, every start-up will have a token as their value accrual model and tokens will be the obvious bootstrapping mechanism for both novel companies and mid-market companies wanting to change their model to allow for community stewardship.

Note: this is not investment advice in any way and you should not buy tokens, either public or private, if you are prohibited from doing so under applicable law or regulation.

Original Publication by Ryan Zurrer • September 3, 2024

What is ReFi?

Regenerative Finance, or ReFi, is an emerging concept that aims to create positive societal impact alongside reasonable financial returns on risk-capital in order to bring ambitious social projects to life that couldn’t exist on philanthropy alone but shouldn’t be co-opted by profit maximization principles. ReFi seeks to be constructive capital to align economic incentives with positive environmental and social outcomes. Usually these models tend to blend non-profit organizations with for-profit corporations in some creative way. I discuss some of the experiments, lessons learned and opportunities moving forward herein.

ReFi Winter

There is a debatable narrative globally that Regenerative Financing doesn’t work and we shouldn’t attempt to innovate at the capital model layer of business. In the last year we’ve seen massive public failures of some prominent experiments in regenerative financing - most notably OpenAI, but also more recently with MAPS/Lykos and the Regenerative Financing Vine that I constructed. Does this mean ReFi is dead or are we just in ReFi winter and needing more innovation and experimentation in this area?

I note that in parallel, crypto has a perception problem. Politicians and TradFi executives often allege that crypto is a giant global grift and an unregulated casino of greedy self-interested shadowy super-coders. While we know that this is largely untrue and most of the space is made up of well-intentioned builders, as a community we need to come together and accelerate the examples of novel business models that program in reciprocity and public benefit directly into the raison d’etre of given projects, thereby using crypto primitives to programmatically unlock value for social good.

Types of Regenerative Financing

De-escalating royalty

One compelling model is the de-escalating royalty, which was a feature of both the OpenAI-Microsoft deal and MAPS-Vine deal. As investors hit milestones on return profile, the percentage of revenues accruing to investors steps down and more of that cash-flow goes back to social mandates (see figure 1 below). Some interesting features of this model are:

1- Investors are paid back plus a reasonable return before capital is spent on non-profit activities.

2- Investors get to participate in the entire upside of the project. This is superior to capped returns that return investment similar to a debt profile. Since many ambitious projects have considerable risk, investors should have a venture-like return.

3- Investors are paid back off of top-line revenue and not bottom-line profit thereby reducing their risk on the efficiency of the organization.

We see in the above figure that ReFi outperforms capped venture returns or even venture-debt over time because investors get to participate in the entire upside of a given business instead of being capped out. I think this is a superior capital model and generally, I would like to see more experiments in this manner.

Buy & Burn

Another model is where tokens are issued but then profits from the entity are used to buy back the assets and burn them with the intention that over time all investors are bought out and the project can return to fully non-profit mandate once the buy & burn is complete. I would like to sincerely thank Nikolai Mushigan for creating the Buy & Burn model in crypto and writing the code for it in MakerDAO. Nikolai represented crypto-native values in everything he did and was a wonderful innovator and friend. We miss him dearly. One issue with this model is tokens become successively more expensive as you traverse up the bonding curve and thus residual token holders could capture outlier value before the project completes its obligations.

Yield farming on a capital asset to use the cash flow for social activities

The advent of yield-farming in DeFi is a really compelling innovation that enables more capital efficiency, which can be useful for non-profit cultural assets such as art. Here a non-profit with valuable assets assumes debt on the capital asset(s) and then yields with the debt. The resultant spread between yield and debt obligation create a cash-flow that is used for a social mandate. This is a model that would work for art institutions and 1OF1 is actively experimenting with art ReFi to create a sustainable art collection. I think this could be an important mechanism to create greater capital efficiency from valuable cultural assets. I do note that because LTVs would typically be conservative in order to protect the underlying asset, this model implies a longer time-horizon to payback and profitability.

Team reciprocity pledges & ecosystem ReFi support

I call on every token-enabled project with governance on the planet, and frankly every fund and company in crypto to have a reciprocity pledge or fund some ReFi projects in their ecosystem. There has never been an industry on planet earth as profitable as programmable money. Some of the extraordinary returns that these innovations generate should support social projects and we would all do ourselves a favour by propagating more reciprocity in our space.

Growth in ReFi

As the ReFi space continues to evolve, several areas show promise for future experimentation:

Carbon Markets or Biodiversity: Developing more sophisticated on-chain carbon credit systems could revolutionize how we tackle climate change. KlimaDAO and Moss.Earth are interesting experiments in this regard.

Elegant 2-Sided UBI with Slashing: It could be possible to implement slashing mechanisms to drive social influence among communities receiving UBI. For example, if a community is expected to maintain a forest in order to receive UBI, slashing could happen if a subset of the community shirks and illegally logs the forest. Socializing the cost of shirking would be an effective way to have community-policing of social outcomes.

Regenerative DAOs: Forming decentralized autonomous organizations focused specifically on funding and managing a social mandate such as stewarding an important cultural asset is an interesting experiment in its infancy. ConstitutionDAO would have been an interesting large-scale experiment in this regard. I personally think Crypto Punks could potentially be stewarded as a community-driven non-profit DAO that seeks to place Punks into museums globally.

Challenges and Considerations

While ReFi holds significant promise, we should be aware of some pain points and potential challenges:

Measurement and Verification: Ensuring accurate measurement of social impact is crucial for the credibility of ReFi projects.

Regulatory Landscape: As ReFi grows, it will need to navigate complex regulatory environments, especially when dealing with real-world assets and questions around whether associated crypto-assets are securities if they are driving a reasonable return for token holders.

Education and Adoption: Increasing awareness and understanding of ReFi concepts among both crypto enthusiasts and the general public is necessary for more experimentation and adoption.

Starting with Reasonable Payback: For a project to hit escape velocity, investors will have to be paid back and then an interesting social mandate executed on. However, projects can’t forget that the weighted average of cost of capital in crypto is high and capital expects a compelling return in this space, even for social projects.

Corporate Structures: Given the challenges that corporate structures imply in mixing philanthropy with for-profit returns, it makes sense that the eventual solution will come from a DAO or on-chain model.

Predictions Moving Forward

Unfortunately, for the next 24+ months we are probably still stuck in ReFi winter. However, compelling experiments are starting to form now that should produce fruit later this decade. As wealth and capital change in the 2030s with the advent of AGI, I do believe that ReFi will become more prominent as we see incredibly profitable organizations form very quickly and cause philosophical questions among their stakeholders.

I firmly believe that blockchain technologies are an innovation that will deliver positive social impact globally. In order to deliver on the dream of Web3 and the decentralization of power and capital on the internet we must continue to fearlessly experiment with mixing for-profit returns with non-profit social contributions. Fortunately, this space is filled with bold visionaries and we have everything necessary to build an inspiring future for the world.

Original Publication by Ogst • March 26, 2024

In this series of blog posts, from the perspective of an institutional yield farming fund, we argue that DeFi should be represented as a knowledge graph to properly assess the risk of the available opportunities. Moreover, we compare this approach with similar work performed in other domains such as cybersecurity. We cover the various uses of such a graph: risk scoring, book correlation, hack contagion, and identifying points of failure.

Glossary

DeFi vocabulary is assumed to be known by the reader.
- Knowledge graph: Structured representation of interconnected data stored as a graph.
- NEO4J : Leading graph database management system.
- Cypher : Querying language for NEO4J, designed for graph data manipulation.

Introduction

For better or worse, for the foreseeable future, the financial use-case is crypto’s killer app. Outside of speculation and decentralized capital formation, the financial use-case is dominated by DeFi. In the realm of DeFi, the concept of “being finance legos” underscores the emergence of intricate structures characterized by numerous interdependencies. These complexities pose significant challenges when assessing risks and understanding protocol dependencies, particularly in interactions with “high-level pools.”

Consider the two-sided Aerodrome USDC/eUSD liquidity pool. This pool consists of USDC, along with eUSD, a stablecoin issued by the Reserve Protocol and collateralized by cUSDC, cUSDT, saUSDC, and saUSDT. These underlying assets are interest-bearing tokens issued by Compound and Aave. Additionally, it’s important to note that Aerodrome itself is a Fork of VelodromeV2, adding another layer of complexity to the risk assessment. Furthermore, the engagement of oracles used by Aave and Compound introduces additional intricacies.

Decomposing these dependencies demands a profound understanding of DeFi. Failure to account for them can lead to oversights in due diligence processes, potentially exposing capital to unforeseen risks.

In summary, navigating the complexities of DeFi requires a comprehensive grasp of protocol dependencies and a meticulous approach to due diligence. By understanding these interconnections, participants can better assess and mitigate risks in the ever-evolving landscape of DeFi.

DeFi as a graph

To alleviate the aforementioned concerns and streamline the process of due diligence and risk scoring, we propose the introduction of the DeFi Graph, which models dependencies within the DeFi ecosystem.

Opting for a knowledge graph structure instead of tables and join operations is primarily motivated by the uncertainty surrounding the length of dependency paths.

For instance, let’s imagine a scenario where a hack occurs on Silo. Now, suppose one needs to check whether a specific Magpie pool could be affected by this hack. Answering such a question would be challenging with a table-oriented database as we would not know how many joins to perform to connect entities. However, it simplifies into a straightforward graph exploration problem when utilizing a knowledge graph structure.

Let us first go through the construction of such a knowledge graph. Then, multiple querying examples will be covered to access the power of that representation.

Graph construction

As mentioned in the introduction, we consider the graph representation problem from the perspective of an institutional yield fund providing liquidity to various DeFi pools, which is Dialectic’s core focus. Thus, we center the graph around liquidity pools. In that knowledge graph, we represent the following nodes and edges to respectively document DeFi entities and relationships linking them :

Note that all these relationships are directed.

It translates into the following schema:

This graph is built automatically using on chain data and NEO4J graph database. NEO4J also allows us to easily introduce new edges to document our crypto native teams knowledge about DeFi. It is important to note that the graph has been specifically generated to cater to the requirements of a yield farming fund focused on stable and ETH-related pools.

Results and implications

The graph is populated based on our internal pool database consisting exclusively of Stablecoins (ex: DAI/USDT), and ETH/LST (ex: WETH/rETH) pools. We obtained a total of 5336 nodes and 14626 edges, detailed below:

The node distribution is primarily dominated by pools, which connect to the rest of the graph through relationships such as CONTAINS, LIVES_ON, and BELONGS_TO. This dominance extends to the edge distribution as well. The prevalence of pool nodes compared to others arises from the presence of multiple pools belonging to the same protocol, residing on the same chain, and containing the same assets.

Now that the graph structure has been properly introduced, one may wonder the identification of the most central nodes within this configuration. To address this query, we define the centrality of a node as the number of pools with a directed path connecting to it. In practice, we run the following formula.

The following distribution is obtained :

As anticipated, we notice that Circle Tether and Maker hold significant influence within the graph, leveraging their respective assets USDC, USDT, and DAI. Additionally, it appears that many pools rely on the Chainlink oracle.

Querying the graph

Now that the graph has been designed, we embark on various querying applications to showcase the effectiveness of this design.

Pool dependencies

The pool-centric nature of the graph allows us the find all the dependencies of a given pool by simply walking down the graph, starting from the pool. In practice we execute the following Cypher query, where “pool1” refers to our internal pool identifier:

In the case of the Aerodrome USDC/eUSD we obtain the following result:

The exploration algorithm initially pinpointed the desired pool and traced paths with edges extending away from the original nodes. This methodology enabled us to break down the tokens within the pool and ascertain their respective issuers. In the case of eUSD, the algorithm also dissected the assets utilized as collateral by the reserve protocol. Furthermore, we identified inter-protocol dependencies, noting Aerodrome as a fork of VelodromeV2. Finally, we flag additional nodes as watchlists, facilitating the monitoring of the protocols.

The exploration algorithm initially pinpointed the desired pool and traced paths with edges extending away from the original nodes. This methodology enabled us to break down the tokens within the pool and ascertain their respective issuers. In the case of eUSD, the algorithm also dissected the assets utilized as collateral by the reserve protocol. Furthermore, we identified inter-protocol dependencies, noting Aerodrome as a fork of VelodromeV2. Finally, we flag additional nodes as watchlists, facilitating the monitoring of the protocols.

Hacks contagion

Based on the graph structure , one can now easily reason on hack contagion. Let us take the following example: A hack happens on Curve.

In a single query we can run a BFS staring from the protocol node representing Curve and return all the touched pool , node that we only follow edges directed toward the failing node.

Running the above query and obtain the following graph:

We first observe a central set of nodes representing Curve itself and all the proprietary pools. One relationship further, we see the protocols Conic, Prisma, and Convex, along with their pools. Additionally, we distinguish the crvUSD asset and pools from other protocols that depend on it.

On that graph one could rank the pools according to their distance to the central Curve not and take reallocation decision based on the results.

N-Order exposures

Understanding the n-order exposures of a yielding book is a complex task, particularly in the management of portfolios comprising numerous strategies. Adopting a superficial “flat” view may lead one to perceive exposures solely to Convex, StakeDAO, and Aura, mistakenly assuming diversification is attained. However, a deeper examination reveals that these protocols all depends on Curve.

Fortunately, such a problem is easily solvable using the DeFi graph. For example, the following function returns the total exposures per protocol, where the allocated_usd property of a pool node represents the capital invested in the corresponding pool.

Following the same logic, one could identify potential points of failure in the portfolio composition. A high dependency on a single bridge could pose a significant risk to the book.

For example, to identify the invested capital exposed to the Wormhole bridge, one can execute the following:

Conclusion

We crafted a graph representation of DeFi and showcased the efficacy of this data format within the realm of a pool’s risk assessment. We delved into the design choices, covering various use cases such as pool dependencies decomposition, n-order exposure unfolding, hacks contagion... We have illustrated that the graph significantly simplifies addressing these issues.

In a forthcoming blog post, we will contrast the designed graph with attack graphs from the cybersecurity domain and introduce a quantitative method to leverage the graph for risk measurement purposes.

Can we learn from investment performance evaluation in traditional finance to help DAOs make better decisions?

Introduction

The literature and practice for investment performance evaluation in traditional finance is extensive. However, the on-chain world has few proxies that could provide DAOs or on-chain investors with the necessary tools to make quantitative investment allocation decisions. There are many ways of conducting portfolio performance attribution analysis, each with its own tradeoffs. However, it all starts and finishes with the benchmark that a portfolio aims to track. In the absence of a benchmark, no performance attribution is really possible.

Download

PDF | EPUB

TLDR:

- DAOs have limited scope to do complex things with respect to allocating their treasuries

- We propose simple methods borrowed from traditional portfolio management to help DAOs make decisions more clearly

- The following frameworks can help communities:

1. Set a benchmark

2. Weigh risk to help select between alternatives

3. Do performance attribution analysis to see if active decisions add or detract value

Other references that could be useful:

Spreadsheet with example calculations

Customizable Dune query with benchmark assets