It is natural to ask when decision markets will get a day of their own — the Saturday Uniswap out-traded Coinbase, the weekend Hyperliquid was the only venue on earth pricing oil, the night Polymarket printed 95% on Trump before midnight.
That is the wrong question. Legibility predicts attention, not adoption. ConstitutionDAO raised $47M from seventeen thousand wallets in under a week, lost its auction live on stream, and dissolved within a fortnight. Its sequel raised 150 times less. Beeple's $69M at Christie's was as visible as a sale can be; art-NFT volume later fell over 90%. Perpetual swaps became most of crypto's derivatives volume without a date anyone remembers. Stablecoins settled $10.9T in 2025 the same way.
The folk theory of breakout moments is a bad way to read this category. The buyers already showed up. The remaining product is standing depth.
Demand is a covenant
A decision market installed by the decision-maker it grades dies of politics. Internal markets at Google, HP, and Ford beat expert forecasts by as much as 25%. Roughly a fifth of Google's employees traded on its internal market. Every one of these programmes died — of politics, law, or neglect, never of inaccuracy. DARPA's policy market was killed by the US Senate within a day of a press conference. Hanson's diagnosis is blunt: markets embarrass powerful people, and powerful people get them killed. a16z's survey of 2025's governance trends does not mention futarchy once.
There is one door that wall cannot close. An organisation born under decision markets has no incumbent to kill them. That is what the MetaDAO cohort is, and it explains where the demand actually showed up: not in governance conversions, but in capital formation. Umbra's ICO drew $155M in commitments against a $3M cap, from more than ten thousand wallets, because the market is the covenant — the team cannot rug a treasury that only a market can spend. MetaDAO's founder now leads with exactly this pitch, and it is the only place demand for the mechanism has ever been oversubscribed.
The monthly decision-market tape does not contradict that. April printed $2.84M; July printed $5.86K. Volume is lumpy rather than trending. Solomon's Gigabus proposal alone was $2.67M in April, so a single large proposal is most of a good month. The July print is one month with nothing large in it, not a measured decline in interest. July resolved one proposal across 13 wallets. Underneath that month, MetaDAO revenue went from roughly $127K in June to roughly $245K in July — +93%. Capital formation works. The decision-making layer is where the thinness lives.
The native record is the product
Organisations born this way already have a control-rights record that no board or token vote can match.
In March, a market approved the liquidation of Ranger Finance — by 6.46% over threshold, on $584K of trading from 57 wallets — and the treasury went back to holders, roughly $5M of it. 01Resolved's forensics found that the wallets carrying the decision were early backers who had lost trust in the team, and that the single largest wallet in the market traded against liquidation and lost. Companies do not fire themselves. This one's owners priced the firing and executed it.
In April, Solomon's biggest proposal — $2.7M traded across 2,276 trades — was pricing toward failure until the team renegotiated with its opponents mid-window, adding a MetaDAO co-founder as a multisig signer and committing to return about $2M of excess capital. The market re-priced the amended deal and passed it. The market was a negotiating counterparty, not a verdict waiting for a camera.
In July, when a four-day-old wallet asked Umbra's DAO for 62% of its treasury, the market refused by 19.55% and billed the attacker $4,623. That refusal landed the same month, on the same chain, as an attacker who bought $4.4M of BONK over a holiday weekend, passed a malicious proposal at 2.9% turnout, and drained about $20M from a treasury whose contracts performed exactly as designed. 01Resolved had made the contrast in December: in a vote, defence is unpaid civic labour, and BonkDAO's defenders were on holiday. In a market, defence is a trade that pays if you are right. July ran the head-to-head.
These are not rehearsals for a later headline. They are the product.
The moat is time multiplied by depth
The clock is real. A decision market runs about 72 hours, 24 of them pre-TWAP and 48 that count; a trade in the final ten minutes carries roughly 0.35% of the verdict. A 48-hour TWAP defeats flash loans. It does not make capital expensive — at prevailing rates, $1M rents for about $1K over two days — and holding a distorted price costs little until opposing capital shows up.
The moat is not time. It is time multiplied by depth.
Umbra's defence was four wallets and $165K that happened to be watching. Its conditional pools held about $52K on a day when $1.5M was in play. A $2M MetaDAO allocation has passed on about $3K of volume. Small is sometimes consensus. Small against a live treasury ask is cheap to push. Both can be true, which is why depth belongs next to every pass edge.
If agentic flow deepens 72-hour conditionals — agents already sit behind more than 30% of Polymarket's wallets — that is a means. The claim is still standing depth for the cohort that already bought the covenant.
An adjacent product, not this desk's climax
Over the eighteen months to mid-2026, prediction markets had a day and compounded it. Kalshi raised at a $22B valuation. ICE, the NYSE's parent, has put more than $1.5B into Polymarket. The Federal Reserve publishes working papers about them.
A deep public conditional on a named merger, defied by its owner, graded by the ordinary tape within a week, captured by whoever owns those rails, is a plausible sequel for that category. It is not a vindication of binding decision markets.
Robin Hanson conceded the point in the paper that invented futarchy: an obeyed market can never be scored against the branch it voided — "we will never know this, and never need to know." Binding books cannot be defied. That is their integrity, and it is a property, not a hole that a later owner is required to fill. Advisory books can be defied, and history says their hosts kill them first: GnosisDAO launched in 2020 promising to govern through futarchy, kept the markets ignorable, and let them fade. The corporate programmes died the same way.
This desk indexes the binding books. A later defiance clip on regulated rails would be news about Kalshi or Polymarket. It is not what would prove the MetaDAO cohort.
The books did not
The covenant sold. The books did not. Standing depth is what these markets still lack. A later Umbra with real books would be this product, working more expensively — not a missing Saturday, and not a day someone else has to volunteer to lose.