Seventy percent. That is the share of the average DAO treasury held in the project's own native token, according to a new report from GSR. The figure landed like a stone in deep water: no emergency proposals, no panic votes, no treasury managers rushing to rebalance. The silence is the story. Most readers will categorize this as a governance footnote or a treasury-management warning. It is neither. It is structural fragility in DeFi's accounting layer.
Let me state the finding precisely. A treasury that holds 70% of its value in its own token is not a treasury. It is a price chart with a governance page attached. GSR calls this a "dangerous feedback loop," which is correct but understated. The danger is not merely that DAOs are overexposed to their own tokens. The danger is that governance architecture makes it structurally impossible to do anything about it before the loop reaches terminal velocity.
DAO treasuries underwrite the entire DeFi ecosystem. They function as quasi-central banks, allocating capital through grants and incentives to developers, liquidity protocols, and security services. They pay salaries, subsidize liquidity mining, and fund the auditors who keep the rails alive. A functional treasury needs outside purchasing power: stablecoins, ETH, assets that can be deployed regardless of the DAO's own token price. When 70% of a treasury is native tokens, that purchasing power is an illusion. The DAO's capacity to fund its ecosystem becomes a direct function of its own market capitalization.
This is self-referential valuation in its purest form. The system produces value, measures value, and consumes value in the same asset. Mapping the topological shifts of a bull run makes the problem invisible: treasury size in USD grows as the token climbs, even with zero real buying power accumulated. The bear market exposes the structure for what it always was.
The mechanics deserve more rigor than the headline. I ran my own stress tests during the 2020 DeFi Summer, deploying $5,000 into Uniswap V2 and Curve to model impermanent loss under high volatility. The durable lesson: a 30% drop in one asset creates damage a 30% recovery does not undo. DAO treasuries carry the same asymmetry. When a native token falls, the treasury suffers a double contraction. The mark-to-market loss is the first axis; the loss of deployable capital is the second. A 70% native basket worth $100 million at peak now holds a fraction of that in spendable assets. Operational budgets for grants, audits, and contributor payments shrink in direct proportion to the token's decline.
The second mechanism is supply overhang. If DAO treasuries are 70% native tokens, real circulating supply is far smaller than nominal supply. Markets price scarcity, not tokens sitting in multisigs waiting for a governance vote. When those reserves move — via a market maker, an OTC desk, or a liquidity program — they enter a market that has internalized scarcity. A 70% treasury is a permanent dam sitting upstream of every price chart. The pressure does not dissolve; it accumulates until a governance vote opens the floodgates.
The third mechanism is governance latency. Even a DAO with the will to diversify cannot execute quickly. The path of least resistance runs through a proposal, a vote, a timelock, and a multisig approval. In a collapse, that path is measured in days, often weeks. My 2024 institutional compliance work taught me that latency is the most dangerous quality a financial system can possess: a system that cannot react to prices only acts after they have moved. DAOs optimized for decentralization have accidentally optimized for slow decision-making. The result is a structural inability to sell.
The loop itself can be described in four stages. Stage one: a macro shock drops the native token 20-30%. Stage two: treasury value shrinks while operating costs in stablecoins do not; the DAO must choose between cutting grants and selling tokens into a falling market. Stage three: on-chain transactions expose the sales, sentiment deteriorates, and spending cuts face political backlash. Stage four: the cycle repeats at a lower price level. GSR's data suggests most DAOs sit between stage one and stage two, with no mechanism to stop the sequence.
That forced selling is the part most treasury models miss. DAOs do not hold native tokens because they want to. They hold them because early distributions — foundation allocations, community rewards, public sale reserves — were denominated in native tokens from day one. When the bear market arrives, these same DAOs must sell tokens at the bottom to pay for auditors, infrastructure, and contributor salaries. This is not portfolio rebalancing; it is a stress-seller with no alternative.
The market has not priced any of this accurately. Most DAO token investors focus on protocol revenue and user growth. Treasury composition is an afterthought, a footnote in the quarterly report. That is why GSR's report matters despite naming no specific DAOs: it reframes a sector-wide weakness that single-project metrics will not reveal until too late. Once institutional risk models incorporate treasury concentration as a discount factor, DAO tokens carry a permanent risk premium.
The concentration risk also propagates downstream. Because DAO treasuries are upstream capital, a large DAO cutting its liquidity incentives does not merely hurt its own token. It pulls the floor from under the protocols relying on that liquidity, the teams relying on those grants, and the users relying on those teams. Ecosystem contagion is not a metaphor here. It is a funding chain composed of single points of failure.
These forces — dual-axis contraction, supply overhang, governance latency, forced selling, and downstream contagion — compound into the systemic risk GSR describes. But I want to correct a misclassification in the report's language. This is not a Ponzi. A Ponzi scheme requires new inflows to pay old participants. The DAO treasury feedback loop requires no new money at all. It runs purely on procyclicality: falling price shrinks the treasury, which reduces funding, weakens sentiment, and feeds further selling. Procyclicality is more treacherous than fraud because it needs no malicious actor. It operates silently through accounting structure alone.
Healthy treasury practice, by contrast, is well established. Institutional treasuries typically hold 30-50% of assets in stable reserves: stablecoins, short-dated bonds, or deep-liquidity collateral. A 70% concentration in an asset whose value depends on the ecosystem it must fund is not risk tolerance. It is a single point of failure wearing an alignment narrative.
GSR's blind spot — and the blind spot in most commentary — is the aggregate itself. The 70% figure says nothing about the quality of the remaining 30%. One DAO may hold a genuinely liquid stablecoin buffer. Another may hold a "diversified" basket of illiquid ecosystem tokens that cannot be sold without moving the market against itself. The concentration ratio is a symptom, not a diagnosis. Without standardized treasury accounting — a shared definition of usable assets — the systemic risk can only be guessed.
Tracing the gas trails of abandoned logic in the treasury reports of dead chains, I suspect we will find that the most dangerous DAOs are not the ones holding 70% native tokens. They are the ones whose diversified 30% is itself a basket of tokens that cannot be liquidated without breaking their own market. Liquidity is not the same as diversity. A reserve basket that cannot be deployed in a week is effectively a 90% concentration.
There is also a regulatory inference hiding in the data. If a DAO's solvency depends almost entirely on its own token price, the argument that token holders are relying on the efforts of others — the fourth prong of the Howey test — becomes harder to dismiss. Trust-minimization is the promise. Terminal dependence on team decisions is the reality. High treasury concentration may eventually appear in securities claims, whether or not GSR intended it.
One more layer. GSR is a market maker, not an academic research unit. Its desks may hold positions in the very tokens its analysts criticize. That does not invalidate the data, but it should calibrate reception. Read this report as a risk signal, not as a prediction.
A fix that arrives late is itself a trigger. When treasury diversification proposals finally pass, markets will price the resulting sell pressure in advance. The cure becomes the catalyst. DAOs that sell native tokens to build stablecoin buffers will find the attempt itself is the confirmation that accelerates the decline.
The next bear market will not ask what a DAO treasury is worth. It will ask how much of that value is real, and how quickly it can be converted when the need arises. Treasury diversification is no longer a best practice. It is a survival prerequisite. The architecture of absence in a dead chain is not the absence of tokens; it is the absence of a credible mechanism to convert governance power into purchasing power. The DAOs that live through the next correction will treat their own tokens as their greatest liability, not their greatest asset. The rest will learn it the way every overleveraged system does: all at once, and in public.