Your treasury is not a balance you glance at; it is money with three jobs that pull against each other. One job is paying contributors next month. Another is surviving two years if revenue dries up. The third is earning enough that the runway stretches — without betting the runway to do it. Most DAOs and small crypto companies do not lose money to a cinematic hack. They lose it to duration mismatch: parking the operating budget in a 12% vault that locks for a month, or leaving two years of cash at 0% in a hot wallet because moving it felt like a project.
Concretely: a 14-person protocol holds $2,000,000 in stablecoins plus a token position. It pays monthly, has committed to two years of grants, and its single largest loss risk is not a contract bug — it is one signer clicking the wrong link on a Sunday. This playbook is written for that team, and for the smaller version of it: the eight-person company that just raised and is deciding where the raise should live. You will get a bucket structure with target percentages, a custody stack a stolen laptop cannot drain, a yield ladder that never puts spendable money at risk, and a monthly cadence that keeps all of it honest.
TL;DR. Run a DAO treasury as three buckets: roughly 40% operating runway in stablecoins earning the ~3.4% onchain lending baseline, roughly 45% in a reserve ladder of tokenized T-bills or RWAs at 4–4.8%, and no more than 15% in an opportunistic sleeve you can afford to lose. Custody is a 2-of-3 or 3-of-5 multisig with hardware keys held by different people — never a hot wallet. On a $2,000,000 treasury that mix blends to about 4.7%, or roughly $94,000 a year. Stop chasing 12% with money you must spend next quarter; that duration mismatch is how treasuries die. Size the split in the portfolio allocator.
Bucket the treasury before you chase a single yield
The order of operations matters: decide how much money does what before you decide which protocol pays the most. Three buckets answer almost every treasury question, because each one solves a different problem. The operating runway keeps the lights on. The reserve ladder earns a real rate on money that has a date attached to it. The opportunistic sleeve lets you take a calculated swing without it touching either of the first two. The percentages below are a starting frame, not a rule — a protocol with 24 months of committed spend should weight the runway higher, one with 6 months and strong revenue can lean into the reserve ladder.
| Bucket | Instrument | Target % | Liquidity | Expected yield | Main risk |
|---|---|---|---|---|---|
| Operating runway (0–12 mo spend) | USDC/USDS in a multisig + money-market or short lending | 40% | Same day (T+0) | ~3.4% lending; 0% if parked in checking | Stablecoin depeg; lending protocol risk |
| Reserve ladder (12–36 mo) | Tokenized T-bills / RWAs (USDY, OUSG, BUIDL-class) | 45% | T+0 to a few business days | 4–4.8% | Issuer/wrapper, eligibility gates, redemption windows |
| Opportunistic sleeve | Curated lending, delta-neutral, audited RWA vaults | 15% | Days to weeks | 6–12% | Smart-contract, strategy and liquidity risk |
Two disciplines hold the frame together. First, size every bucket by what you would do if it fell 10% overnight: the runway must be sized so a stablecoin wobble never stops payroll, and the sleeve must be sized so a total loss is survivable. Second, keep the buckets physically distinct — a separate multisig Safe per bucket, or at least clearly labelled sub-accounts. When the money for next month's salaries lives in the same address as your degen experiment, one rushed approval can spend the wrong pile. The full risk arithmetic for the two lower-risk buckets is laid out in tokenized treasuries vs DeFi lending.
Worked example: allocating a $2,000,000 treasury
Put the frame to numbers. Assume $2,000,000 in stablecoins, 21 months of runway at current spend, and no debt. The runway bucket holds 40% ($800,000) in a same-day stablecoin position earning the roughly 3.4% onchain lending baseline. The reserve ladder holds 45% ($900,000) in tokenized T-bills at an illustrative 4.4% net. The opportunistic sleeve holds 15% ($300,000) at an illustrative 9%. Ranked by yield, weighted by size, the mix blends to about 4.7% a year.
| Bucket | Amount | Yield | Annual income |
|---|---|---|---|
| Operating runway | $800,000 | 3.4% | $27,200 |
| Reserve ladder | $900,000 | 4.4% | $39,600 |
| Opportunistic sleeve | $300,000 | 9.0% | $27,000 |
| Blended | $2,000,000 | ≈4.7% | $93,800 |
Two things stand out. The reserve ladder earns the most dollars despite being the quietest bucket — a 1-point rate advantage on the largest allocation beats a flashy sleeve every time. And the sleeve, the riskiest 15%, contributes about a quarter of the income but none of the payroll safety. That is the correct trade: put the risk where a loss is survivable, and keep the rate competition in the bucket that has no deadline. Model your own version with the portfolio allocator before you move a dollar.
The custody stack: multisig, separate hands, and a signing policy
Bucket design is worthless if one key can move everything. A treasury needs a multisig with a threshold above half the signers, so no single compromised device is enough, and below all of them, so one lost device does not lock the funds forever. In practice that means 2-of-3 as the floor and 3-of-5 once the treasury clears roughly $1–2 million. Every signer key is a hardware wallet held by a different person, and no two seeds live under the same roof; a fire or burglary should never collect a controlling set in one visit.
The part most teams skip is the written signing policy. Put it in the governance docs: which signers approve what, at what amounts, and how a spend moves from proposal to execution. A common structure is a small standing allowance any single signer can draw for routine costs, capped by an on-chain spending-limit module, with everything above it requiring the full threshold. That single control keeps day-to-day operations from becoming a two-building ceremony while keeping the strategic stack behind two signatures. Our 30-day field journal of running a portfolio behind a 2-of-3 explains the mechanics and the friction in detail: 2-of-3 Safe multisig for 30 days.
The reason a single hot wallet is disqualifying is not hypothetical. In February 2025 attackers moved roughly $1.46 billion out of a Safe controlled by Bybit without touching the contract — they poisoned the off-chain signing interface, served a transaction that looked routine, and let valid signatures execute a hidden payload. A hot wallet is the same failure mode with a lower bar: one extension, one blind signature, no second screen to catch it. Hardware keys do not render full calldata either, which is why the control that matters is verifying the same transaction on a second, independent interface before anyone confirms.
Keys worth the ceremony. A multisig is only as strong as its signers' devices, and the whole point of separate hands is that each held key is genuinely hard to steal. DifiCalc earns a commission through the link below at no extra cost to you; the recommendation above stands on its own and is separate from that commission — see our affiliate disclosure.
The yield ladder: 0% checking, 3.4% lending, 4–4.8% T-bills
Think of treasury yield as a ladder, not a race. The bottom rung pays nothing: cash sitting in a "checking" position — an idle hot wallet or an uninvested exchange balance — earns 0% and exists only for immediate, unscheduled spend. The next rung is onchain stablecoin lending, which has settled around a 3.4% baseline through 2026 for blue-chip markets; this is where the operating runway belongs, liquid the same day. The third rung is tokenized T-bills and short-duration RWAs at roughly 4–4.8%, a step up in rate that costs a redemption window and an issuer relationship. The 30-to-50 basis points between rungs is the price of duration, and it is worth paying only for money that has no near-term deadline.
The rung most teams want to skip to is the one that pays 12% and up. Do not. A rate that far above the stablecoin baseline is not free yield; it is compensation for risk a treasury usually should not be paid to take — smart-contract exposure, a lock-up that collides with payroll, or a strategy that unwinds badly in the exact stress that makes you need cash. If the money is spoken for next quarter, the correct rung is the low one. If you have genuinely surplus capital and can tolerate a loss, the opportunistic sleeve is where that risk belongs and nowhere else. Product-level mechanics for the reserve ladder, including KYC and redemption, are in the step-by-step guide to buying tokenized treasuries, and the surrounding rate environment is tracked in best stablecoin yield 2026.
Monthly reporting, review cadence, and the four controls that prevent losses
Yield compounds quietly; drift compounds too, and only a cadence catches it. Run a monthly treasury close: record the balance in each bucket, the yield actually earned versus what you projected, and every position opened, closed or resized that month. Then rebalance when a bucket drifts more than a few points from target — trim the sleeve when it has run, top up the runway when spend is ahead of plan. A monthly rhythm is also your early-warning system for depegs, a stalled redemption or an approval you forgot to revoke. The routine is spelled out in the monthly DeFi portfolio review.
Four controls, applied together, neutralize the losses that actually happen to treasuries. Signer collusion: keep the threshold above half and split custody across people and places, so no pair of insiders can move funds and no single location holds a controlling set. Phishing a single signer: require hardware keys, cap what any one signer can execute alone, and verify each transaction on a second independent interface before confirming. Unaudited vaults: only deposit into contracts with public audits that you have actually read, and hard-cap any single position so one failure never touches the runway. Unreviewed approvals: grant narrow, limited allowances, and revoke them on the monthly close instead of letting a dormant unlimited approval sit as an open door. None of these is exotic; the discipline is doing all four every month, not the month you feel like it.
Sources and further reading
- Safe {Wallet} — documentation — multisig owner architecture, thresholds, spending limits and the propose-confirm-execute flow.
- ethereum.org — security — wallet hygiene, key management and the threat model behind multi-key custody.
- rwa.xyz — live dashboard of tokenized Treasury and short-duration RWA products, yields and holders.
- BlackRock — BUIDL — official fund terms, holdings and distributions for a flagship tokenized T-bill product.
Frequently asked questions
What percentage of a DAO treasury should be in stablecoins versus tokenized T-bills?
A workable 2026 baseline is roughly 40% in stablecoins for operating runway (0 to 12 months of spend), 45% in a reserve ladder of tokenized T-bills or RWAs for money you will not touch for 12 to 36 months, and no more than 15% in an opportunistic sleeve you can afford to lose. The split tracks duration, not conviction: the shorter the horizon, the more the money should sit in stablecoins at about 3.4% rather than in a 4–4.8% T-bill wrapper with a redemption window.
How many multisig signers does a DAO treasury need?
Use a 2-of-3 multisig as the floor and a 3-of-5 for larger treasuries. The threshold must sit above half the signer count so no single compromised key can move funds, and it must be below all of the signers so one lost device does not brick the wallet. Hardware keys should be held by separate people, ideally in separate locations, with none of them stored next to another signer's key.
Why not chase 12% yield with treasury funds?
Because yield is payment for taking risk you can or cannot afford. Money you must spend next quarter needs same-day liquidity, so a 12% strategy that locks funds, relies on an unaudited vault, or carries depeg and liquidation risk creates a duration mismatch: your liabilities are short and your assets are not. Any rate far above the roughly 3.4% onchain stablecoin baseline or the 4–4.8% tokenized T-bill range is compensation for risk a treasury should usually not be paid to take.
How often should a DAO review its treasury?
Run a full review monthly and rebalance when a bucket drifts more than a few percentage points from target. The monthly close should record balances by bucket, yield actually earned versus expected, and any position opened, closed or grown. A monthly cadence catches drift, depegs and stale approvals before they compound, and it produces the paper trail that keeps signers accountable.
Is a single hot wallet ever acceptable for a DAO treasury?
No. A single hot wallet means one browser extension, one malware infection or one blind signature can drain the treasury, and it offers no second signature to catch a spoofed transaction. The February 2025 Bybit incident, where roughly $1.46 billion was moved after attackers swapped the implementation in a poisoned signing interface, was not a multisig failure; it was a reminder that one signer reading one poisoned screen is enough when there is no second check. Keep only small, capped operating balances in a hot key behind a spending-limit module.
Size your three buckets in minutes
Split any treasury across runway, reserve and sleeve, then see the blended yield and the risk each bucket carries before you sign.
Open the Portfolio AllocatorRelated: lock down custody with the 2-of-3 Safe multisig journal; compare the two reserve instruments in tokenized treasuries vs DeFi lending; follow the walkthrough to buy tokenized treasuries; and run the routine in the monthly DeFi portfolio review. Manage the whole mix in the portfolio allocator.
This article is education, not financial, legal or accounting advice. Rates and product terms change; verify every figure before you move treasury funds, and size positions so that no single loss stops payroll.