Solana fee reform backed by Yakovenko fails despite majority

Solana’s Aug. 28 governance vote on SGP-0003 ended in rejection despite a larger number of tokens voting For than Against. The final tally was 142.844 million SOL For, 50.146 million SOL Against and 72.025 million SOL Abstain. Because abstentions were included in the two‑thirds approval denominator, the proposal fell about 33.83 million SOL short of the supermajority required.

The governance snapshot recorded 265.015 million SOL participating, equal to 61.14% of the 433.486 million SOL snapshot, so quorum was met. The Constitution and the governance FAQ count For, Against and Abstain toward participation and the two‑thirds denominator. The frozen ballot text excluded abstentions, but the official system applied the ratified rules when finalizing the result.

SGP-0003 proposed changes to transaction fees and a staged resource-fee ramp. The plan would replace Solana’s fixed 5,000‑lamport fee per signature with a 2,500‑lamport inclusion fee per transaction paid to block leaders, add a resource fee charged by requested scheduler cost that would be burned in full, and keep priority fees flowing to leaders. The ballot set a three‑stage path for the resource-fee rate: one‑tenth of a lamport per requested cost unit initially, then one‑quarter and finally one‑half of a lamport.

Anatoly Yakovenko publicly backed the initial one‑tenth rate before the vote and argued that validator revenue encourages more people to stake. After the vote he proposed splitting the package into a narrower proposal to replace the fixed signature fee and a separate proposal on whether validators or an automatic mechanism should set future rates.

Charging by requested resources would affect applications unevenly. Transactions that request large compute limits or are inefficient could face higher costs or be rejected for insufficient balance, while low‑resource transactions could pay less. Legacy validator vote transactions would need updates to compute‑budget handling and fast‑path logic before the planned Alpenglow activation. Every validator client would need the relevant feature gates enabled before the consensus‑breaking change could activate. The technical design for distributional rules is encoded in SIMD-0553 and requires separate implementation steps.

Positions on the vote cut across operators and delegated‑stake holders. Named supporters included Figment, Staking Facilities, Kiln and P2P.org. Named opponents included Jupiter, Drift, Bitwise Onchain Solutions and Forward Industries. Jupiter held about 11.78 million SOL; the For side required roughly 33.83 million additional SOL to reach the two‑thirds threshold.

Under Solana’s governance process, an SGP records a directional stake mandate while SIMD proposals and client releases handle technical design and deployment. Finalizing an SGP locks the tally; implementation follows as a separate engineering and coordination process. Holders of 15% of active stake retain the ability to force a network vote under the rules that govern fundamental economic changes.