The 5% and 95% Problem: Ethena's Buyback Proposal Has a Governance Bug
The governance post was published, and the market responded the way markets do: ENA jumped 27% in two days. But the post contains a contradiction that no one seems to have reconciled. The milestone table says the buyback starts at 5% of protocol revenue. The text below it says 95% of foundation net income. Both numbers appear in the same document. Neither is crossed out. This is not a rounding error. It is a governance bug that will surface the moment the first buyback is executed.
Let me be precise about what Ethena actually is, because the narrative has outpaced the mechanics. Ethena runs USDe, a $4.22 billion synthetic dollar. The protocol holds spot assets and shorts an equivalent amount of perpetual futures. This is a delta-neutral basis trade. The yield comes from funding rates โ the periodic payments long positions make to short positions in perpetual markets. When funding is positive, Ethena earns. When it flips negative, the protocol bleeds. The strategy is not new; quantitative desks have run basis trades for years. What Ethena did was tokenize the position and sell it as a stablecoin alternative. That packaging is the innovation. The underlying trade is as old as derivatives.
The buyback proposal is a fee switch. The smart contract functionality is trivial โ route a percentage of protocol revenue to a buyback mechanism. The fee switch itself is a simple conditional redirect that checks the current USDe supply against a threshold and routes the corresponding percentage. The complexity is not in the Solidity. It is in the economic assumptions baked into the threshold schedule. The design question is not technical. It is economic. The proposal ties the buyback rate to USDe supply: 5% at $7.5 billion, scaling to 25% at $25 billion. The trigger condition is the first problem. USDe sits at $4.22 billion. The buyback does not activate until the supply grows 78%. That is not a near-term event. It is a conditional promise that may take multiple quarters to reach, if it is reached at all. The market has priced the buyback as if it is imminent. It is not.
Here is where the analysis gets uncomfortable. The governance post specifies two different revenue pools. The milestone table references a percentage of protocol revenue. The accompanying text commits to 95% of foundation net income. These are not the same number. Protocol revenue is gross. Foundation net income is after expenses, after operational costs, after whatever the foundation decides to deduct. A 5% cut of gross revenue and a 95% cut of net income could produce wildly different buyback amounts. The post never reconciles them. This is either sloppy drafting or deliberate ambiguity. Both possibilities are bad. Sloppy drafting suggests the governance process lacks rigor. Deliberate ambiguity suggests the foundation wants flexibility to choose the interpretation that suits it later. Neither outcome is favorable to token holders who voted for the proposal based on what they believed it promised.
I have audited enough governance proposals to know that this pattern repeats. The market reads the headline โ "Ethena announces buyback" โ and prices the token up. The details are left for later. The details are where value is actually distributed. Math doesn't lie, but governance documents can. The 5% versus 95% discrepancy is not a footnote. It is the entire economic substance of the proposal.
The governance participation makes this worse. The Snapshot vote, as of Tuesday evening, showed 17.8 million ENA in favor, zero against, across 87 votes. ENA's total supply is roughly 15 billion tokens. That is a participation rate of approximately 0.1%. Eighty-seven wallets decided the direction of a $4.22 billion protocol's token economics. This is not community governance. It is a quorum of convenience. The foundation proposed, a handful of large holders approved, and the market extrapolated enthusiasm from a vote that almost no one participated in. Trust is a vulnerability, not a virtue. In this case, the market is extending trust to a governance process that has not earned it.
The comparison to Hyperliquid is instructive. Hyperliquid already executes daily automatic buybacks using nearly all of its trading fees. The mechanism is live. It is not a proposal awaiting a vote. It is not gated behind a supply threshold. It is running. Binance has a longer track record with quarterly BNB burns. Ethena's plan, by contrast, is conditional, ambiguous, and unexecuted. The market is treating all three as equivalent signals. They are not. One is a promise. Two are operating mechanisms.
The regulatory backdrop matters here. The article notes that in the United States, paying token holders is no longer treated as legally risky, and value-return plans have become standard practice within roughly a month. That is a meaningful shift. If the SEC has softened its stance on token buybacks, it removes a compliance overhang that previously discouraged this structure. But it does not resolve the underlying question of whether USDe itself constitutes a security. The Howey test factors โ investment of money, common enterprise, expectation of profits, reliance on the efforts of others โ all point in uncomfortable directions. A more favorable regulatory environment for buybacks does not immunize the underlying asset from securities classification. It just makes the buyback itself less legally fraught.
The structural risk is not in the smart contract. It is in the market that generates the revenue. Ethena's yield depends on perpetual funding rates remaining positive. That is a function of market structure โ the balance between long and short demand in perpetual markets โ not of Ethena's code. In a prolonged bear market or a low-volatility regime, funding rates compress or go negative. The protocol's revenue dries up. The buyback, even if activated, has nothing to fund it. The entire model is a bet on the persistence of positive funding. That is a market bet dressed up as a yield product.
There is also the centralization dependency. Ethena's basis trade executes on centralized exchanges. The protocol's revenue stream depends on the health, solvency, and cooperation of those venues. A single exchange failure, a withdrawal freeze, or a change in fee schedules directly impacts the protocol's ability to generate the revenue that the buyback is supposed to distribute. This is not a theoretical concern. It is the operational reality of running a delta-neutral strategy at scale.
The market has already moved. ENA trades near $0.158, up roughly 77% in 30 days. The buyback narrative is priced in. What is not priced in is the execution risk โ the 5% versus 95% ambiguity, the 78% supply growth trigger, the 0.1% governance participation, and the funding rate dependency. When the vote concludes on September 2, the market will get a binary outcome: the proposal passes or it fails. But the real information will come after, when the foundation clarifies which revenue pool the buyback actually draws from. That clarification will determine whether this was a value-return event or a narrative event.
Privacy is a protocol, not a policy. The same logic applies to buybacks. A buyback is not a policy statement. It is a protocol โ a set of rules that execute deterministically, without foundation discretion, without ambiguity. Ethena's proposal is not yet a protocol. It is a policy with two conflicting definitions of its own terms. Until that is resolved, the buyback is a story the market is telling itself, not a mechanism that is running.
The vote will pass. The ambiguity will remain. Watch the foundation's post-vote announcement. That is where the actual economics will be decided. The 5% and the 95% cannot both be true. The market will find out which one is, and the price will adjust accordingly.