A bootstrap budget is the money a DAO spends before it is able to decide anything. Someone has to pay for the legal entity, the first contributors, the audits and the infrastructure, and that spending starts before there is a functioning vote to authorise it – so in practice a foundation, a development company or an elected transition body sets the first budget, and the DAO's first budget vote is held afterwards. What that vote is actually deciding varies, and the difference is worth naming: it can be a genuine appropriation, a renegotiation of terms, or a ratification of transfers that already happened.
This is the mirror image of council dissolution. Both are moments where authority sits with a body the token holders did not elect through the DAO's own process – at the start because the process does not exist yet, at the end because it is being wound down. Both are also where the gap between formal governance rights and actual control over funds is widest.
Three constraints push the budget ahead of the vote, and none of them is avoidable by better intentions.
The consequence is that the DAO's leverage over its own first budget is much narrower than its formal voting rights suggest. It rarely gets to decide whether the money is spent. It gets to decide what conditions attach to the remainder.
The clearest worked example is Arbitrum DAO's first proposal. AIP-1 was put up in March 2023 as an omnibus that, among other things, asked token holders to ratify the Arbitrum Foundation's setup – including an "Administrative Budget Wallet" already holding 7.5% of the ARB supply. Delegates read it as being asked to approve a transfer that had already occurred, and voted it down: the Snapshot record closes at 100,343,537 against to 16,418,340 for, with 14,120,336 abstaining across 26,141 voting addresses – roughly 77% against.
What happened next is the part worth studying, because it shows exactly what the rejection did and did not achieve. A follow-up proposal asked the Foundation to return 700M ARB to the DAO treasury outright. That also failed, and by a wider margin: 118,334,390 against to 20,526,942 for across 43,392 addresses, about 84% against. The same electorate that refused to bless the allocation also refused to claw it back.
What passed was the renegotiation. AIP-1.1 disclosed that 0.5% of supply had already been transferred out of the budget wallet, put the remaining 7% (700M ARB) under a four-year linear lockup running from the date of the Snapshot approval – about 175M ARB unlocking per year – and published a first-year operating budget of $36M, split $16M general and administrative, $9M R&D, $5M technical infrastructure and $6M events, marketing and communications, to be funded 50% in USD and 50% in ARB. It carried 137,012,025 for to 1,224,386 against across 45,230 addresses, roughly 98% in favour. A companion governance proposal, AIP-1.2, passed with a similar margin.
Read the four votes as a sequence and the pattern is unambiguous. Ratification failed, reversal failed, and terms-plus-disclosure passed overwhelmingly. The DAO's real instrument was never the yes/no on the allocation. It was the lockup schedule, the reporting requirement, and the itemised budget it could measure the next request against.
Uniswap did it the other way round. The Uniswap Foundation was put to a temperature check in August 2022 and its funding approved by the DAO before it began operating – $74M in total, split roughly $52.3M in grants and $21.7M in operations, sized to run through the end of 2024. An ex ante appropriation, which is the textbook answer to the Arbitrum problem.
It still did not produce a fixed budget, for a reason specific to DAOs: the request was denominated in the DAO's own token. The first tranche was written as 2,547,002 UNI targeting $20M, and by the time it settled on 25 August 2022 UNI had fallen 13.7% to $7.05, so the Foundation received $17.3M rather than $20M. A year later it returned to ask for the remainder, $62.37M – the $56.7M outstanding plus a 10% buffer against exactly that volatility – and that request went to an on-chain vote in October 2023.
The lesson generalises past this one case. A budget voted in the DAO's own token is a budget whose real size is set by the market between the vote and the transfer, and the correction is always another vote. That is one of the mechanical reasons a DAO holding only its own token has less budgetary control than its balance sheet implies – the reason treasury diversification is a governance question and not just a portfolio one.
A third shape inverts the ordering again: here the bootstrap grant is explicitly small, and the treasury transfer is conditional on the DAO existing first. In Foundation Update: Moving to Maintenance Mode, published 28 April 2026, the Radix Foundation said it had approved a two-stage grant of $67,000 and 10m XRD to fund the DAO's set-up and initial operation, with the community-elected Accountability Council leading the entity formation, and stated that "when a community entity is established, the intention is for the remaining treasury to transfer to it, subject to the legal and compliance steps any such transfer requires." It also named what happens if that does not land: "if a community entity is not established within the Foundation's operational timeline, assets and social logins will be held dormant where possible."
That post is dated deliberately here, because the status it describes is the kind that decays without anyone publishing a correction. Re-checked on 3 September 2026, it remains the Foundation's most recent statement on the transfer, so the arrangement below should be read as the position on 28 April and not as a live confirmation.
Sequenced that way, the pre-governance spend is a rounding error against the treasury it precedes, which removes the Arbitrum-style ratification problem almost entirely. It creates a different one. Until the transfer completes, the DAO governs a set-up grant while the assets it was formed to steward sit with the outgoing body, and the transfer itself is gated on legal steps no vote controls. A DAO in that position has full formal authority over almost none of its eventual balance sheet, which is a live risk rather than a settled arrangement – and one worth re-checking against the primary source rather than assuming, since a stalled transfer looks identical to a pending one right up until it doesn't complete.
The entity the transfer waits on is being formed as Radix DLT DAO LLC in the Marshall Islands, and the ratification of the DAO's charter and policies closed its discussion phase on 18 September 2026, with a Temperature Check next. radix.wiki tracks both on its Radix DAO transition board.
Across all three shapes the token holders' effective levers are the same short list, and none of them is the headline vote.
What the DAO generally cannot recover is money already moved. Arbitrum's electorate tried and declined by a larger margin than the original rejection. Treating the first budget as reversible is the mistake; treating it as the last unconditional grant the founding body will receive is the accurate reading.
A caper has no pre-governance phase to budget for, because it has no launch round to fund. Its treasury fills from the bonding curve as people trade, from the first buy onward, so there is never a stretch in which money has been raised and governance is not yet running. Every trade pays a share into the treasury. The founder is paid out of the same trades as they happen, and that share goes to whoever holds the caper's founder badge, so there is no founding allocation to ratify later. See raising funds for how capital arrives.
Every spend from a caper's treasury goes through a proposal its holders vote on, from the first one; the only other way money leaves is a member leaving with their share. There is no unconditional first tranche and nothing spent before the vote existed. The cost is that a caper cannot spend anything before there are holders to authorise it.