Solana is scheduled to replace its consensus system on a live network on October 21, 2026. Anza, the team behind the Agave validator client, said Alpenglow "goes live on mainnet-beta Wednesday, October 21, ~14:50 UTC" and that "finality drops from seconds to ~150ms," according to Anza's announcement on X. The current system, TowerBFT, takes roughly 12.8 seconds to finalize a transaction, according to CoinMarketCap Academy.
Most coverage focuses on the speed. For analysts, the more useful question is what finality actually measures, why it matters for settlement, and which claims can be checked once the switch happens. That is the focus of this piece.
What Finality Means
A blockchain transaction goes through stages. First it is included in a block. Then the network's validators vote on that block. Finality is the point at which the block cannot be reversed under the protocol's rules, as long as enough validators follow those rules.
This is different from throughput, which counts how many transactions a network can process each second, and from block time, which is how often new blocks appear. A network can produce blocks quickly and still need many seconds before those blocks are final. For anyone using a chain to settle trades, collateral, or payments, finality is the number that decides when a transfer can be treated as complete.
How This Analysis Was Built
This piece uses three types of information:
1. Protocol facts from the developers, including Anza's announcement and Solana Foundation material as reported by trade outlets.
2. Test results from Solana's community test cluster, as reported by Solana Compass.
3. A market snapshot of SOL taken at 02:38 UTC on October 19, 2026, from Perplexity Finance, used only to translate a validator fee into dollars.
Alpenglow has not yet run on Solana mainnet. Every performance number below is either a design target or a test result. Neither is a measured mainnet outcome.
What Changes Under Alpenglow
Alpenglow has two main parts, according to a Solana developer guide. Votor handles voting and finality. Rotor handles how new blocks are spread across validators. Only Votor is scheduled for October 21. Rotor does not yet have an activation date, according to Bitcoin.com News.
The biggest practical change is how validators vote. Under TowerBFT, validators submit their votes as on-chain transactions. Under Votor, validators exchange signed votes directly and combine them into compact certificates instead of writing every vote to the chain, according to The Coin Republic's summary of Solana Foundation guidance. Removing vote transactions from blocks frees space that votes used to take up.
Votor has two paths to finality, according to Solana Compass:
1. Fast path. If a block receives votes from at least 80% of staked SOL in the first round, it is final right away. The target is about 100 milliseconds.
2. Slower path. If a block receives at least 60% but less than 80% in the first round, a second round begins. Another 60% vote makes it final. The target is about 150 milliseconds.
The two paths run at the same time, so the network can still finalize blocks when some validators are slow or offline.
How the Switch Happens
The migration is designed to happen without a network restart. The upgrade is turned on at the start of epoch 1064. After that, there is a short handoff window in which blocks carry only validator votes rather than user transactions. Votor takes over once a block gets votes from at least 82% of staked SOL, according to Altcoin Buzz. Bitcoin.com News put the handoff window at about 18 minutes.
Validators have several tasks before the switch. They must run a client built on Agave 4.3, register a new signing key on their vote account, and hold enough SOL to pay a new Validator Admission Ticket fee. The Solana Foundation set that fee at 0.8 SOL per epoch, according to The Coin Republic. Anza made Agave 4.3.2 the recommended release for mainnet validators on October 16, citing Alpenglow fixes, according to Solana Compass.
The Worked Example
Here is what the published numbers mean in plain terms.
Finality time. Moving from about 12.8 seconds to about 150 milliseconds means finality would be roughly 85 times faster. To get that figure, divide 12.8 seconds by 0.15 seconds. In practice, a transfer that today needs around 13 seconds to be treated as settled would need less than a fifth of a second, if the target holds on mainnet.
Test results. On the community test cluster, slots averaged 214 milliseconds and 96% of blocks finalized through the fast path, according to Solana Compass. That suggests most blocks reached the 80% vote level in the first round in testing. Mainnet has more validators, more varied hardware, and real economic activity, so the share could be different.
Validator fee. At a SOL price of $104.89 in the October 19 snapshot from Perplexity Finance, the 0.8 SOL admission fee works out to about $83.91 per epoch for each validator. To get that figure, multiply 0.8 by the SOL price. The dollar cost will move with the SOL price. Whether the fee is material depends on each validator's stake and income, which this piece does not measure.
Validator set size. Alpenglow caps the consensus set at 2,000 validators as a design constant, according to Solana Compass.
Other Upgrades on the Calendar
Alpenglow is one of several network upgrades this month. Each targets a different part of the infrastructure.
1. Ethereum. The Glamsterdam upgrade is scheduled for the Hoodi test network on October 26, 2026, where the gas limit will also rise to 200 million, according to the Ethereum Foundation. Glamsterdam focuses on block building and gas pricing rather than finality. Its mainnet date has not been decided.
2. Zcash. The NU7 upgrade went live on Zcash's test network on October 4 and cuts the target time between blocks from 75 seconds to 25 seconds, according to BigGo Finance. The same report said developers planned to set a mainnet activation height on October 20.
The common thread is that networks are competing on how quickly and cheaply they can confirm activity. The approaches differ. Solana is changing consensus itself, while Ethereum is focused on block capacity and fee rules.
What the Data Does and Does Not Tell Us
The data supports a narrow set of claims. Alpenglow is designed to cut Solana's finality from about 12.8 seconds to about 150 milliseconds. Test results show the fast path working for most blocks in a controlled setting. The switch is scheduled for October 21, and validators face new software, key, and fee requirements.
The data does not show how Alpenglow will perform on mainnet. Test clusters are smaller and calmer than a live network. It also does not show that faster finality will lead to more usage, more fees, or any change in the SOL price. Finality is one part of settlement. Users and institutions also care about uptime, fee stability, and how a network behaves under heavy load. Finally, the change in validator requirements could affect how many validators take part, but no data is yet available on that.
Why This Matters for Analysts and Institutions
Finality is likely to become a more common point of comparison between networks. Analysts can prepare by tracking a small set of measurable items after the switch:
1. Median and worst-case finality on mainnet, compared with the 150 millisecond target.
2. The share of blocks finalized through the fast path, compared with the 96% test result.
3. The number of active validators and the share of stake taking part in voting.
4. Block space freed by removing vote transactions, and whether it fills with user activity.
5. Any outages or stalls during or after the handoff window.
Those numbers will show whether the change works as designed. Claims about faster settlement should be judged against them rather than against the design targets alone.
Explore Presolt
Presolt brings network, token, and market data into one research workflow, so analysts can compare infrastructure changes across chains with consistent metrics. Explore Presolt to follow the networks you track.
Compliance Disclaimer
Presolt provides data and editorial for informational purposes only. Nothing here is investment advice or a recommendation to buy or sell any asset. Figures reflect developer statements, test results, point-in-time market snapshots, and third-party reports, and may change. Past behavior does not predict future results. Crypto assets are volatile and you can lose your capital.