Infrastructure · Governance
Solana's First Real Vote Came Down to Two Firms Changing Their Minds
The network's inaugural binding governance vote cleared a two-thirds bar by roughly a third of a percentage point. How that margin was assembled is a more interesting story than what passed.
Solana ran its first network-wide binding governance vote this week, and the headline result is that one of the three proposals on the ballot squeaked through. That framing undersells it. The interesting artifact here is not which proposal passed — it is the discovery of exactly how few decisions it takes to move a supermajority on a network this large.
Voting opened August 22 and closed at the end of epoch 1023, around 15:30 UTC on August 27, per Bitcoin.com News. Results were confirmed and reported the following day. Three proposals ran independently; a rejection of one had no bearing on the others.
What was actually on the ballot
SGP-0001 established a Solana Constitution — a formal governance framework that weights voting power by economic stake while letting token holders override their validators' votes. It passed comfortably, with about 86% support, according to Crypto Briefing. This is the one that matters most in the long run, and it drew the least attention, which is usually how constitutional questions go.
SGP-0002, the "Double Disinflation" proposal implemented through the technical change SIMD-0550, raises the annual rate at which new SOL issuance declines from 15% to 30%. It is not an immediate cut; it makes the existing taper run twice as fast, reaching the network's terminal 1.5% rate in roughly 2.8 years instead of about 5.7. Its authors estimate that removes some 18.9 million SOL from six-year issuance projections. This is the one that passed by a hair.
SGP-0003 would have restructured transaction fees through SIMD-0553, replacing the flat 5,000-lamport fee with a 2,500-lamport inclusion fee paid to validators plus a separate usage-based resource fee that is burned entirely. It failed to reach two thirds.
The arithmetic of a razor
SGP-0002 needed two thirds of participating stake — 66.67%. It got 67.00%.
SGP-0002 final tally, as a share of participating stake
Two mechanical details produced that margin, and both deserve more scrutiny than a passing vote usually gets.
The first is that abstentions counted toward the denominator. The 20.63 million SOL that formally abstained were included in "participating stake," which means abstaining was not neutral — it functionally worked against passage by making the two-thirds bar harder to clear. Voters who wanted to register presence without taking a side were, arithmetically, voting no. Whether every abstaining operator understood that is not something the tally can tell us.
The second is the inverse. Kraken had voted against SGP-0002 and then withdrew its vote near the close, per Crypto Briefing; Galaxy moved from abstaining to supporting. Withdrawing a "no" is materially different from abstaining: it removes that stake from the denominator entirely, rather than parking it in a bucket that still counts. Galaxy's move did both jobs at once — it subtracted from the abstain pile and added to the yes pile. On a vote decided by under a million SOL, two institutional decisions in the closing hours were the whole ballgame.
The concentration was visible from the start
None of this should have been a surprise, because the qualifying round showed the same shape. To reach a formal vote, a proposal first had to gather signalled support from 15% of staked SOL — about 65.16 million. Reporting on that phase from Solana Compass in early August found that although more than 70 validators had signalled support, roughly two thirds of the stake behind it came from a single entity — Helius — which had also authored the proposals.
That is not an accusation of anything improper. Stake-weighted governance does exactly what it says: it weights by stake, and stake is concentrated in professional operators because running validators at scale is a business. But it does mean the phrase "the validators voted" is doing a lot of concealing. A more accurate sentence is that a small number of large staking businesses voted, and everyone else's stake mostly followed the operator it was delegated to. SGP-0001's override provision, which lets delegators countermand their validator, exists precisely because someone anticipated this — and it will be worth watching whether anyone uses it.
Passing a vote is not shipping a change
One last thing the coverage tended to compress: SGP-0002 is a governance mandate, not a live protocol change. The technical implementation, SIMD-0550, still has to clear validator client updates and a mainnet activation before anything about issuance actually moves. Governance approval is the instruction; the code is the outcome. Those have come apart before on other networks, and the gap between them is where the real engineering risk lives.
The Take
The result is less notable than the discovery. A major network just ran its first binding vote and learned that its supermajority threshold is within reach of two firms adjusting their positions before the bell. That is not a scandal — it is an accurate measurement, and it is far more useful than a comfortable margin would have been. The right response is not outrage about Kraken or Galaxy, who voted their stake as the rules permit. It is to notice that a two-thirds bar means very little when the top of the stake distribution can move it, and that the abstain option quietly behaves as a "no." Governance systems reveal themselves under load. Solana just ran its first load test and got a legible result. Design the next one knowing what this one showed.