The problem: a treasury full of its own token

Most DAOs are rich on paper and poor in practice. The treasury holds a large block of the DAO's own governance token and very little of anything a contributor can be paid in. Selling that block on the open market is the obvious fix and the worst one: the float is thin, the sale is visible, and the price the DAO realises falls as it sells. So DAOs reach for two instruments that move supply without touching the order book.

An OTC placement is an outright sale of treasury tokens to named counterparties at a negotiated price, settled bilaterally. A market-maker loan lends tokens to a trading firm so it has inventory to quote with, and is repaid in tokens or in cash at the firm's choice — which makes the loan a call option the DAO has written. Both are governed by a public vote, which is why the terms of deals that would be confidential anywhere else are readable in full on DAO forums.

Neither instrument has a home in the DAO's own accounting vocabulary. A placement looks like revenue in the quarter it closes and like dilution for years afterwards; a loan looks like an asset that has left the building. Both are tokenomics decisions dressed as operations.

The market-maker loan, and the option hidden inside it

The structure is standard enough that Lido's own request for proposals listed it as the expected shape: when Lido asked market makers to bid in February 2022, it told them to state "all relevant parameters. (Possibly: treasury loan, embedded call options, length.)" That RFP followed a four-option sentiment check that closed with 57% of votes for opening a public call rather than negotiating privately — the DAO deciding, first, that the terms should be argued in the open.

The clearest public example predates it. In September 2021 Index Coop passed IIP-83, lending 100,000 INDEX — 1% of supply — to Wintermute Trading for one year. Repayment, per the published terms, was Wintermute's choice of three:

• the full 100,000 INDEX back; or
• $5 million in stablecoins or ETH, if an Index product or INDEX itself reached a Tier-1 centralised exchange with Wintermute's help; or
• $7.5 million otherwise.

Read it as an option and the economics are plain. Wintermute returns cash whenever the tokens are worth more than the cash, and returns tokens whenever they are worth less. The DAO has sold a one-year call on 1% of its supply, struck at an effective $50 or $75 per INDEX. A forum member did that arithmetic during the vote and objected that the DAO was "selling a circa $1.5m option for zero value"; another walked through the payoff explicitly — if INDEX traded at $250 a year later with no listing, Wintermute would repay $7.5 million and keep $25 million of tokens — and called it a "heads I win, tails Index loses" structure. The proposal passed on Snapshot with 127,634 votes for and 19,281 against, roughly 87% in favour.

The premium was never separately priced, and that is the general defect rather than a quirk of one deal: the loan is negotiated as a service contract, so the option rides along inside it and no line item ever names it. The counterparty side is candid about the components when asked. A market maker proposing to quote stSOL on Solana told the Lido forum that "normally, MMs would ask for a fee as well as the 'inventory' (assets) to quote the token, but I only wanted the inventory" — fee and inventory as two separable asks, one of which most DAOs never see itemised.

Index Coop's relationship with Wintermute ran to term and was not renewed on the same footing. In December 2022 the DAO reopened the search, noting it was "past our agreement term end dates" and wanted a fresh look now that larger institutional partners were in play — the ordinary end of a service-provider relationship, renewed by expiry rather than by review.

The OTC placement, and Lido's premium

A placement is simpler: the DAO sells a block of treasury tokens to buyers it names, at a price it fixes, for stablecoins it can spend. The received wisdom is that such blocks trade at a discount, because the buyer is absorbing size and accepting a lockup. Lido's largest placement did the opposite.

In July 2022, in the depths of the bear market, Treasury Diversification #2 proposed selling 20 million LDO — 2% of supply — at a 7-day TWAP plus a 50% premium, a price of $1.452153 per LDO, for $29,043,051.43 in DAI. Dragonfly Liquid took 10 million and other strategic participants the remaining 10 million. The DAO's stated goal was two years of operating runway in stablecoins, and the premium was justified as giving new partners "skin in the game" while blunting the incentive to flip.

The term that matters most is not the price. It is this one: "Acquired tokens will be unlocked. Full voting power will be granted immediately upon closing of the sale." The buyers paid above market for a block that carried governance rights from day one, with nothing vested and nothing cliffed. That is the exact inverse of the norm the same ecosystem applies to its own contributors, where multi-year vesting with a one-year cliff is treated as the minimum evidence of alignment. Control was delivered instantly; only the founders' and contributors' claims are made to wait.

The premium is legible once you see what was actually sold. A discount compensates a buyer for illiquidity and price risk. A premium is what a buyer pays when the thing acquired is not just exposure but a governed position in a protocol whose direction they intend to influence — a price for the vote, not for the token. The 105-post thread that followed argued the point at length, and the deal closed.

How to read a deal before voting on it

Both instruments are legitimate. What makes them dangerous is that their cost is structurally invisible in the format DAOs vote in. Four questions extract it:

• What is the option worth? A loan repayable in tokens or cash at the borrower's election is a written call. Price it — strike, size, tenor, implied volatility — and put the number in the proposal. If nobody in the DAO can price it, the DAO is not equipped to sell it.
• What is the effective strike relative to spot? IIP-83's $50 and $75 were far above INDEX's price at the time, which is what made the structure defensible to most voters. A near-the-money strike on 1% of supply is a very different trade wearing the same clothes.
• Does the block vote, and from when? Tokens sold unlocked with immediate voting power change the electorate, not just the balance sheet. A 2% placement to a handful of aligned funds is a governance event.
• What happens at term? Both Index Coop deals show the same failure mode as service-provider contracts generally: they end by expiry, not by an evaluation against the KPIs the original proposal listed. Write the review into the deal or it will not happen.

The instruments also age. Lido's treasury, having sold 2% of supply for runway in 2022, was by August 2025 debating the opposite trade — a dynamic buyback programme directing 70% of incoming liquid tokens to LDO buybacks against a $145M liquid balance, with thresholds pausing buybacks below $50M. A treasury that placed tokens at a premium in a bear market and bought them back from a surplus three years later has, on net, run a discretionary trading book. That is a real strategy; it is rarely the one the DAO voted for.

How Caper approaches this

A caper does not have the problem these instruments solve, because it never accumulates a block of its own token that it then has to place. Buying in happens on the caper's bonding curve: tokens are minted along the curve as capital arrives, and the same closed-form curve prices the sell leg, so there is a standing bid and ask without a counterparty being hired to provide one. The contract's compute_xrd_received integrates the curve between two supply points to price a sale, and the configured trade fee — 0.5% at instantiation, and rewritable by update_fees like the other three fees — is what the protocol charges for the round trip. There is no inventory to lend because the curve is the inventory.

The exit right closes the other half. Any member can call exit at any time, and the share of the treasury they receive is their canonical vote weight (t·v)/(V·T) — the same compute_vote_weight the protocol uses for voting, computed pre-burn from the governance tokens and the soulbound vote tokens together, both of which the call asserts are non-empty and consumes. So a member does not need a market maker to get out, and a large holder cannot enlarge their claim by buying tokens alone. (Verified against contracts/logic/src/lib.rs.)

What a caper gives up is real and worth stating. There is no mechanism to sell a strategic block to a named partner, no way to bring an institution onto the cap table on negotiated terms, and no lever to buy listings or deepen a centralised-exchange book — the things Index Coop and Lido were actually buying. A caper trades the ability to do those deals for never having to price an option it does not know how to price.

References

• Index Coop — IIP-83: Wintermute Market Making Proposal (6 Sept 2021), and the updated proposal thread with full terms.
• Snapshot — IIP-83 vote result (index-coop.eth): 127,633.67 FOR / 19,281.50 AGAINST.
• Index Coop — Index Coop Invites Market Makers To Partnership Talks (5 Dec 2022).
• Lido Research — Market makers and CEX Listings (2 Feb 2022), the four-option sentiment check.
• Lido Research — Request for Proposals: Market Making (23 Feb 2022).
• Lido Research — Treasury Diversification #2 (18 July 2022), 105 posts.
• Lido Research — Lido on Solana: liquidity on Openbook and Jupiter (13 Dec 2022), a market maker stating its own terms.
• Lido Research — Dynamic Buyback Program for LDO (7 Aug 2025).