Cardano's path to its Dijkstra upgrade ran through a governance vote that almost did not clear. Stake pool operators approved the renewal of the network's constitutional committee with 51.18% of voting weight. The threshold was 51%. The whole Dijkstra roadmap sat on 0.18 of a percentage point.
That margin decided more than one committee's membership. The constitutional committee reviews governance actions before they take effect, and four of its seven seats expired at the end of epoch 653 on September 6. Had the vote failed, the committee would have dropped to three members, below the minimum of five, and four categories of governance would have frozen: treasury withdrawals, protocol parameter changes, adopting a new constitution, and hard fork initiations.
Hard fork initiation is the one that matters here, because Dijkstra is a hard fork, and Cardano has run this play before through its earlier Van Rossem hard fork.
The stake pool vote nearly missed
Delegated representatives cleared their bar more comfortably, at 69.36% against a 67% requirement. Stake pools were the problem. Billions of ADA sat in non-voting stake, and idle weight counts against passage under Cardano's rules, so pools that never cast a ballot nearly stalled governance by inaction alone. Intersect, the group that coordinates the network's governance tooling, had flagged stake pool turnout as the decisive factor before the deadline.
SPO approval rested just 0.18 percentage points above the line after billions of ADA in non-voting stake weighed against passage.
The renewal seated four new members, among them Marek Mahut and the Eastern Cardano Council, with terms running to epoch 799. Turnout gates like this recur across chains that put upgrades to a vote. Zcash met a version of it when its NU7 upgrade needed a coin-weighted threshold before it could ship.
Node 11.1.1 shipped before the ink dried
Days after the committee renewed, Cardano released node 11.1.1, the first of four planned versions leading into the Dijkstra fork. It strips out legacy tracing code and cleans up memory problems left over from 11.1.0. This is early groundwork, and the features arrive in stages.
| Release | Role | Timing |
|---|---|---|
| Node 11.1.1 | Legacy cleanup, memory fixes | Shipped |
| Node 11.2 | Most features, no Leios | About a month out |
| Node 11.3 | Hard fork release candidate | One to two months after 11.2 |
| Node 12.0 | Software that flips the era | Not scheduled |
Node 11.2 is expected about a month after Intersect's September 5 update and will carry most of Dijkstra's changes for testing, minus the Leios consensus work. A public testnet called DijkstraNet follows, exercising Plutus V4 and nested transactions, while a separate network, MusashiNet, puts Leios through its paces. Node 11.3 becomes the release candidate later, and version 12.0 moves the chain into the new era.
When the fork could land
Intersect has published two activation windows instead of a single date. The nearer one, at moderate confidence, spans December 5, 2026 to January 4, 2027. The later one, at high confidence, runs February 24 to March 26, 2027. Neither is locked. Both depend on testing that has not happened, and on approval from the same committee that just squeaked back into quorum. Solana is walking a parallel line with its Alpenglow upgrade and a validator vote of its own.
What Dijkstra brings is a move to protocol version 12: a new ledger era, nested transactions that let one transaction carry others inside it, the Plutus V4 contract environment, and Linear Leios, which threads parallel processing around the existing Praos consensus.
The engineering will move on its own cadence of node releases and testnets. The open variable is turnout. Governance friction is not unique to Cardano, as Arbitrum's recent move against three grant teams showed on another chain, and the next test for the Dijkstra timeline comes when stake pools vote on the fork itself, under the rules that nearly tripped them this time.