DAO tokenomics is the design of a decentralized organization's native token: how many exist, how they enter circulation, who receives them, and what holding one entitles you to. In a DAO the token usually does double duty – it is both a claim on the organization's economic upside and the instrument of control over it – so its design decides, at once, who profits and who governs. Get it wrong and the two goals collide: a token distributed to maximise a launch valuation can hand control to short-term speculators, while one distributed to align long-term stewards can starve the treasury of the capital it needs. a16z's DAO canon treats token design as the foundational choice every other governance decision inherits.
The first lever is the supply schedule – the total number of tokens and the rate at which they reach circulation.
Where the supply goes at launch shapes governance more than any later vote. Common allocations:
The recurring failure is concentration. However carefully a token is distributed, holdings tend to pool: studies of large DAOs repeatedly find voting power concentrated in well under 1% of addresses, so a token that is nominally "one token, one vote" is in practice governed by a handful of whales and delegates. Concentration is the root cause of most of the governance attacks catalogued elsewhere in this wiki, and the problem token-weighted voting inherits by construction.
A reserve for ecosystem incentives can be tested before launch. A case study by Kim, Laskowski, Zargham and colleagues describes how Insolar, an 80-person Swiss blockchain company, worked with researchers from York University and the engineering firm BlockScience to simulate a pool of tokens paid to application developers on an exponentially decaying schedule, running ten simulated years many times over for four starting pool sizes. A larger starting pool left the foundation's treasury worth more at the end, although it then held fewer tokens, while developers' long-run income came out much the same whatever the pool's size. The foundation gained from the gap between fee revenue that grew and subsidy payments that shrank.
Because a governance token bundles value and control, its holders can face a conflict: the profit-maximising choice and the protocol-maximising choice are not always the same, and a holder with no lasting stake can vote for the former. Vitalik Buterin's Moving beyond coin voting governance sharpens the point: a voter who borrows tokens to vote has "zero net exposure" to the outcome – full control, no downside – which is the mechanical root of vote-buying and the bribe markets that grew up around governance. Much of modern tokenomics is an attempt to re-couple the two, so that whoever holds the votes also holds the risk.
Vote-escrow (ve) models are the best-known answer. Pioneered by Curve's veCRV, they require holders to lock their tokens for a fixed term – up to four years – to receive voting power (and a boosted share of rewards), with power scaling by lock length and decaying as the lock runs down. Locking forcibly ties control to a multi-year commitment: you cannot vote and immediately sell. The design succeeded so well it triggered the "Curve Wars," in which protocols competed to accumulate veCRV to direct Curve's emissions – and spawned an open bribe market where locked voters are paid for their votes, re-introducing the very vote-buying it aimed to suppress. The lesson generalizes: no purely token-based scheme fully escapes the value-versus-control tension; it can only move where the leak appears.
Early design advice treated each extra role for a token as a strength. William Mougayar's 2017 business guide scored a token one point for each of 20 roles it could play, from product access and governance to payment, security deposit and profit distribution, on the reasoning that more roles held a business model in place more firmly. Later analysis argued that the roles conflict: holders save a token they expect to appreciate rather than spend it, so the investment role and the payment role undercut each other (see token velocity).
Gnosis separated the roles from the start. Its January 2017 token model had holders lock GNO, the token it sold, to generate a second token, each unit good for $1 of platform fees, with issuance adjusted to keep the supply of these fee credits at 20 times average monthly use. Holding GNO worked like holding a licence, a one-time purchase that paid for access to the platform, while the fee token had no speculative element and could double as stable collateral in prediction markets.
Fees paid in any other token were to be auctioned for GNO, with each GNO collected permanently lowering the platform fee. The fee token was announced as WIZ and deployed on Ethereum mainnet as OWL. Matt Lockyer proposed a similar split in 2018 for a hypothetical dispute-resolution protocol: an investment token sold to fund the project, a separate token bought to vote on disputes, and the DAI stablecoin for the funds held in escrow, so that the token users spend is kept apart from speculation on the project.
What a governance token is worth is genuinely contested, because much of what it confers – the right to vote – has no direct cash flow. Where a token also carries economic rights (a claim on fees, a buyback, or a treasury share), analysts value it on those flows; where it is "governance-only," its price rests on the expected future value of control, which is far harder to pin down and a live subject of academic work – see, e.g., Designing a Token Economy: Incentives, Governance, and Tokenomics (2026). This is the token-valuation problem in the economics section, and it is why the sharpest tokenomics designs try to give the token a concrete economic claim rather than leaving it a pure governance right.
A caper takes a deliberately different route to nearly every choice above. There is no emissions schedule, no vesting cliffs, and no allocation table to negotiate: a caper's tokens are issued by a bonding curve, sold out of a fixed supply of 100 billion that is minted once, when the caper is created, into the caper's own inventory vault with no role able to mint more or burn any, and returned to that vault when someone sells. So "supply in circulation" is simply whatever the market holds – not a figure set by a launch spreadsheet.
The founder's allocation is not a reserved lump but half of a skim on every trade, buy or sell, that starts at 100% of a trade on an empty curve, falls to 1% by the time 10.8% of the supply has sold, and halves again with every further tenth of it. The skim divides three ways: 49.5% to the caper's own treasury in XRD; 50% to the founder, 90% of that in XRD and 10% in the caper's tokens; and 0.5% to the protocol-wide $CAPER treasury in tokens. The XRD part is priced on the XRD the trade moves through the curve, and the token part is a slice of the tokens the trade moves. The rate is bounded and legible rather than hidden in a vesting schedule, but because sales pay it too, the total a founder collects has no fixed cap as a share of supply.
On the value-versus-control tension, a caper does not lean on locking. Instead of ve-style time-locks, voting weight combines the stake a member holds with the participation they have actually shown – so influence accrues to members who have actually voted – the record is minted 1 per ballot cast and on nothing else, since trades stopped minting it on 11 September 2026 – and a holder who has never cast a ballot earns no say at all, however the position arrived.
Crucially, that same earned weight sets each member's claim on the treasury at exit – and that claim is the weight itself, tokens held combined with votes cast, not a pro-rata slice of tokens alone: voice and exit are the same number, which re-couples control to genuine economic exposure without a bribe-prone lock market. The token's economic backing is concrete rather than speculative – it is redeemable against a real treasury – which sidesteps the "governance-only, hard-to-value" problem at its root. The exact mechanics live on the linked markets and governance pages; this article is the map of the wider landscape they sit in.