Governance ReviewEpochs 647 to 6492 Aug 2026 to 17 Aug 2026Mainnet
The committee cleared two withdrawals that the DReps did not fund
Scalus and Bifrost cleared the Constitutional Committee, six of its seven members each, and neither secured enough DRep support, at 51.8% and 33.8% against a 67% bar. In the same window the DReps approved ₳120M for Cardano PRIME and agreed to raise the treasury spending limit by ₳150M.
Two bodies, two questions
Scalus asked the treasury for ₳2.5M and Bifrost for ₳12.3M, both for development work. Six of the seven committee members voted yes on each and one abstained on each, which is 100% of the ballots that count, well above the two thirds a withdrawal needs there. Neither secured the 67% of counted DRep power an approval needs: Scalus reached 51.8% and Bifrost 33.8%. The vote on both ended at the boundary to epoch 647, and every body has to clear its bar at the same boundary, so the money stayed where it was.
That gap is not a contradiction. Under the rules of Cardano governance the Constitutional Committee votes on whether an action is constitutional, while the DReps vote on whether it should happen, which for a treasury withdrawal means whether the treasury should pay for it. The committee’s yes does not guarantee funding, and a DRep no does not necessarily imply a constitutional objection.
Two that failed on both counts, and one that was paid
Global Order Book missed both bars at once, at 26.7% of DRep power and 42.9% in the committee, where three voted yes and four no. dOSPO and OMF ran out one epoch later on 2.4%, the lowest DRep share of the seven decided here, and on 57.1% in the committee, where four voted yes, none against, and three cast no ballot at all.
One request came through. Daedalus asked for ₳1.8M, the smallest of the six amounts here, took 77.1% of DRep power and six committee votes, and was enacted at the boundary to epoch 647.
DRep support of six withdrawals, epochs 647 to 649
In percent of counted DRep power, abstaining power left out, against the 67% bar
The treasury gets ₳150M more room
Four requests expired, and the DReps widened the envelope the next ones will be judged against. The net change limit for epochs 613 to 713, the cap on what withdrawals may take out of the treasury across that run, went from ₳350M to ₳500M, agreed in Net Change Limit with 62.4% of counted DRep power behind it.
By the end of epoch 649, withdrawals since epoch 613 had used ₳337.4M, almost the whole of the old ₳350M allowance, and all but ₳1.8M of that was drawn before this window. The increase leaves ₳162.6M for the epochs still to come.
A ceiling is not a payment, and every withdrawal still needs its own approval. Cardano PRIME got that approval in this window, on 80.1% of DRep power and six committee votes, ratified at the boundary to epoch 649 with its ₳120M due one epoch later, outside this window. The limit itself is set by an info action, which the ledger neither ratifies nor executes, so the ceiling is an agreement recorded by a vote rather than a rule the chain enforces. DRepTalk tracks what has been drawn against it on the treasury overview.
Where the abstentions went
Four of the ten largest DReps abstained on every one of the four withdrawals that ran out: Yoroi W₳llet at ₳568.9M, Blockdaemon at ₳333.6M, EMURGO at ₳297.2M and Sebastien Guillemot セバ at ₳183.7M.
An abstention is not the same as staying away. DRep power that casts no ballot is folded into the no side, while abstaining power is left out of the count altogether, so those ballots added no yes power and took their weight out of the calculation, where silence would have left it against the proposal. Three of the four voted yes on Cardano PRIME, and Blockdaemon abstained there too, on all seven actions decided in this window.
The numbers underneath
At the base of all this, fewer DReps are voting at all. The count that cast at least one ballot in the last 12 epochs fell from 313 at epoch 647 to 302 at epoch 649, against 342 at epoch 643, and epoch 649 itself saw 74 DRep ballots against 722 in epoch 643. Across all three bodies 341 ballots were cast in the window, replacement votes included. In the window 109 DReps cast a final ballot, each counted once.
The power behind those votes is also more concentrated. The power delegated to DReps barely moved, from ₳5.27B to ₳5.28B, a figure that leaves out the always abstain and always no confidence options, which are delegation targets rather than voters. What moved is where it sits: the ten largest DReps held 49.9% of it at epoch 646 and 51.6% at epoch 649, the most concentrated reading in a record reaching back to epoch 508.
The ten largest DReps, epochs 627 to 649
In percent of delegated DRep power, the two default options excluded
What comes next was already being voted on. The renewal of the Constitutional Committee, Update Constitutional Committee 2026, and a change to minPoolCost and the Plutus memory limits, Reduce minPoolCost to 75 ada and increase Plutus Memory Limits (Part 2), both had voting windows ending at the start of epoch 653. Three more were open at the close, Name the Protocol Version 12 hard fork “von Bergen“ to epoch 651 and the withdrawals Reimburse Ikigai Info Governance Action Deposit. and Governance Incentives Framework 2026 to epoch 656. The tallies of all five belong to the windows that follow this one.
Also decided
| Action | Outcome | DRep yes |
|---|---|---|
| Se7en Labs: Daedalus Wallet Maintenance and Improvements 2026-2027 | Enacted 647 | 77.1% |
| Net Change Limit: Cardano Treasury (Epochs 613-713) | Closed 647 | 62.4% |
| Global Order Book connect Cardano DeFi to increase transaction | Expired 647 | 26.7% |
| Withdraw 120,000,000 ada for AlphaGrowth’s Cardano PRIME | Ratified 649 | 80.1% |
The numbers behind the window
- Delegated to DReps, at the start of epoch 647
- ₳5.27B
- Delegated to DReps, at the start of epoch 649
- ₳5.28B
- Votes cast in the window, superseded votes included
- 341
- DReps whose final ballot fell in the window
- 109
- Treasury, at the start of epoch 647
- ₳1,446.4M
- Treasury, at the start of epoch 649
- ₳1,453.7M
- DRep shares are the share of counted power: abstaining power is left out, and power that cast no ballot is folded into the no side. Committee shares count members, with abstentions left out and members who cast no ballot counted against.
- The committee minimum size in force at the end of epoch 649 is not stated. The data behind this review holds no protocol parameter snapshot at or before this window.
- Votes cast is every ballot of epochs 647 to 649 across all three roles, superseded votes included.
- How many actions were open for a vote in each earlier epoch is not part of this record, so the fall in ballots is a reading, not an explanation of it.
- An info action carries no amount on chain, so the net change limit for epochs 613 to 713 and the ceiling it replaced come from the treasury registry this site maintains. What has been drawn against it is summed from the enacted withdrawals themselves, as of the end of epoch 649. That registry also names the action that closed in this window as the one that raised the ceiling.
- Five actions were still inside their voting windows when epoch 649 closed. They are named without tallies, because the record holds only their tallies as they stand today, which is after this window.
- Every lowest and highest reading named here is bounded by this record, which starts at epoch 508.
- This edition was written after the window closed. Apart from the payment of the ₳120M withdrawal at the boundary to epoch 650, which the text names as outside the window, it carries no outcome reached after epoch 649.
Data behind this edition: a snapshot of the site's governance data taken 2026-09-08 06:00 UTC, frozen as epochs-647-649.pack.json next to the article. Every figure above traces to a field in that file. Its git blob hash 63ea2687be32 pins the exact bytes, whatever later happens to the branch.