EL7.AI
CryptoMediumUpdated×2•Originally published 21 October 2026•Updated 21 October 2026•
4 min read

Solana activates Alpenglow, targeting transaction finality of 150 milliseconds

Key Facts

1The Solana Foundation targets transaction finality of about 150 milliseconds, compared with 12.8 seconds previously.
2The validator admission ticket in Alpenglow costs 0.8 SOL per epoch.
3The admitted validator set includes a maximum of 2,000 eligible validators, with priority given to larger stakes.
4Collection of the first admission fee begins at the start of epoch 1064 to cover participation in epoch 1065.

The Alpenglow upgrade went live on Solana's mainnet-beta on 21 October 2026, according to an announcement by Anza, targeting transaction finality of about 150 milliseconds. Anza said the transition took place at slot 459,653,000, around 14:00 UTC, after activation was still pending that morning. The Solana Foundation compares the target with the previous finality time of 12.8 seconds under the Tower consensus system. The change concerns when a transaction becomes permanently committed to the network's ledger, affecting transaction and payment settlement. Transaction execution and user fees remain unchanged with the transition to the new system.

The new voting mechanism, called Votor, uses two paths to finalize blocks depending on the proportion of participating stake, according to Anza's technical introduction to the protocol. When 80% of stake participates, the system can finalize a block through a single voting round, the path the company describes in its design. When only 60% of stake responds, block finalization requires two voting rounds instead of one. Anza's documentation explains that the transaction statuses "confirmed" and "finalized" have become equivalent under Alpenglow, both meaning that a transaction is permanently committed to the ledger. The protocol's bootstrap certificate also includes an aggregated cryptographic signature, alongside a map identifying the validators covered by that signature.

The transition also changes the conditions for validator participation, with Anza's documentation setting a recurring admission fee of 0.8 SOL per epoch. The fee is burned from each admitted validator's vote account, according to the operational documentation for vote accounts. Alpenglow's admitted participant set is capped at 2,000 eligible validators, according to the system's published rules. Validators with larger stakes receive priority when this set is selected, making stake a factor in admission. For stakers, Galaxy said the upgrade does not alter how staking rewards are earned, explaining: "Nothing changes in how stakes earn rewards."

For applications and users, Triton One said transaction submission and interaction with existing programs on the network continue in the same way after the upgrade. Priority fees, tips and account data also remain unchanged as the consensus system changes, according to the company. Alpenglow's timing target concerns transaction finality, while block production and transaction execution remain separate processes from that measure. CoinDesk, meanwhile, said the reported total transaction count could fall after activation because validator votes will disappear from blocks. That could happen even as users continue trading and making payments at the previous pace, according to its explanation of the effect of removing voting from the block ledger.

The mainnet launch was preceded by a testing program that included attack simulations, penetration tests and repeated transition rehearsals, alongside a security competition, according to Anza. Galaxy described its operational experience, saying it had participated by running a validator on the Alpenglow testnet since June 2026. During that participation, the company witnessed 6 testnet rebuilds from a new genesis block. Its experience also included a transition attempt that failed across the entire testnet, according to its account of the testing phase. That attempt was followed by a rollback of the transition, among the events Galaxy recorded while operating the validator before the mainnet launch.

Anza published its original introduction to Alpenglow on 19 May 2025, presenting the consensus system design that later went live on Solana. According to that introduction, the protocol was designed to tolerate adversarial stake of up to 20% of the stake participating in the system. The design adds the ability to tolerate unresponsive stake of 20%, on top of the adversarial stake specified by the company. These conditions concern the behavior of consensus participants, whether they act adversarially or stop responding while the network is operating. In the same design, Anza linked the number of voting rounds to the amount of responsive stake, allowing blocks to be finalized through different participation paths.

The Solana Foundation sets the start of epoch 1064 as the time for collecting the first validator admission fee under the transition to Alpenglow. That fee covers participation in epoch 1065, according to the foundation's published transition schedule. For subsequent development, the foundation plans to introduce the Rotor protocol after Votor's adoption, in a later release of the upgrade. Rotor will replace Turbine, the protocol used to propagate block data, according to the announced plan for Alpenglow's components. The foundation has not set a release date for Rotor, placing it only at a later stage after adoption of the new voting mechanism.