Four of seven seats expire at the end of epoch 653, and below five members the committee cannot ratify treasury withdrawals, parameter changes or the hard fork initiation that Dijkstra depends on.
Four of Cardano's seven constitutional committee members lose their seats at the end of epoch 653, and the governance action meant to replace them has not cleared either of the two thresholds it requires. The chain will keep producing blocks. It will simply lose the ability to approve anything.
The action, formally Update Constitutional Committee 2026, was submitted in epoch 646 and needs simultaneous approval from two separate groups: 67% from delegated representatives, who vote on behalf of ADA holders, and 51% from stake pool operators, the entities running the network's block-producing pools. Intersect, the member-led organisation that administers Cardano's governance process, published a vote record on 31 August showing DReps at 66.3% and SPOs at 39%. One group is within a point of the bar. The other is not close.
The deadline itself has been widely misreported, which is its own small indictment of how legible this system is. Several outlets named 1 September as the cutoff. Cardano's own governance notice puts the expiry at 21:44 UTC on 6 September, at the end of epoch 653 rather than its beginning. Epochs run exactly five days, and epoch 651 opened at 21:44:51 UTC on 22 August, which places the start of epoch 653 on 1 September and its close five days later. The community notice is unambiguous about which of those two dates matters.
If the action expires unratified, the committee drops from seven active members to three, two short of committeeMinSize, the protocol parameter that sets the floor at five. Under CIP-1694, an undersized committee cannot ratify any governance action that requires its approval, and that list is the important part: treasury withdrawals, protocol parameter updates, changes to the constitution, and hard fork initiation. Blocks would still be produced. Transactions would still confirm. Staking rewards would still be paid. What stops is Cardano's capacity to spend its own money or change its own rules.
That lands badly against the upgrade calendar. The Dijkstra era is planned in two phases: the first introduces the Dijkstra ledger era and ships Ouroboros Linear Leios as a fully activated feature, targeted for the fourth quarter of 2026, and the second activates Ouroboros Peras through an intra-era hard fork. Both phases need governance actions the committee has to sign off. Intersect has warned that a break in continuity could slow the path to Dijkstra, which is a careful way of saying the engineering can finish on schedule and still sit in a queue.
The situation is recoverable, and it is worth being precise about why. Intersect has said Info actions and Update Committee actions remain available while the committee is undersized, so a fresh appointment action can refill the seats. CIP-1694's ratification table also exempts motions of no confidence from committee approval. There is no permanent lock here. There is just a delay measured in epochs, during which the treasury cannot move and the parameters cannot shift.
The more interesting failure is the one underneath the numbers. Cardano's design lets stake sit in an "always abstain" position, which removes it from the active calculation entirely rather than counting against a proposal. The drepActivity parameter compounds it: set at 20 epochs, roughly 100 days, it means a representative who stops voting has their delegated power go dormant, disenfranchising holders who did nothing wrong and may not know it happened. The result is a system in which apathy and opposition produce the same outcome, and apathy is cheaper. Cardano is not being blocked by a faction. It is being blocked by absence.
Comparisons are unflattering. Cardano cancelled its own 2026 Summit earlier this year after a treasury vote missed its threshold, so this is a repeat rather than a first. Other chains route around the problem with softer defaults; Solana's validators cleared a contentious emissions question in the burn vote that closed last month, and DAO votes elsewhere have managed decisive outcomes on smaller treasuries, including Seamless's proposal to freeze its own contracts permanently. The distinction is not that Cardano's voters are less engaged. It is that Cardano requires more of them to show up, and gives the absent ones a way to look like they never existed.
The published figures should be read with some care. A CardanoScan reading cited on 25 August put DReps at 41.7% and SPOs at 12.0%, while a GovTool snapshot on 17 August showed 32.46% and 1.95%. Those tools compute role-specific active stake differently, so subtracting one from another to produce a rate of progress is not sound arithmetic. The direction is clear enough without it, and so is the shape: DRep participation has climbed steadily, SPO participation started near zero and remains the binding constraint.
ADA traded near $0.20 through the week, down roughly a tenth over seven days on CryptoSlate's data, having drawn attention earlier in the year when six months of CME futures put it into the SEC's 75-day ETF lane. Price is a poor instrument for reading governance risk, and nobody should pretend a governance stall is what moved it.
Cardano spent years building the most elaborate on-chain governance apparatus in the industry, with formal thresholds, a written constitution and an elected committee. It is four days from being stopped by nobody voting.
