Cut down to 0.15 seconds from 12.8 seconds—Solana wants to reduce “transaction finality” time by 98.8%, and this change has only just started running for real for the first time.
📢 群里有实时分析
On September 26, Solana co-founder Anatoly Yakovenko reposted the news about Alpenglow landing on devnet, with just one line: “You're not ready for this.” Anza’s engineers activated this new consensus at epoch 1167, followed by a countdown—then block production was taken over by the new version. Previously, it had already switched successfully on testnet—which means Alpenglow is now running simultaneously on two public test environments.
What it replaces is Solana’s TowerBFT, which it has used for years. The old logic had validators write votes to the chain—voting while accumulating consensus. Votor, the new approach, moves vote coordination off-chain, then compresses the result into a cryptographic certificate to send back. Votes no longer crowd out block space; the saved capacity is given back to user transactions, and validator operators’ voting costs drop as well. The fault-tolerance assumptions are: up to 20% of staked funds behaving maliciously, and another 20% going offline at the same time—the network still needs to keep moving.
What’s really being rewritten, though, is the “waiting feeling.” Solana now takes about 12.8 seconds for a transaction to reach finality. Alpenglow’s goal is around 150 milliseconds; when validator responses are fast enough, the fast path can be pushed down to about 100 milliseconds. At 150 milliseconds, that’s roughly an 85× improvement.
Put it alongside the price: SOL is currently around $121.4, up about 8.6% over the week. It has reclaimed the $120 level. The market’s watched support range is $115 to $120, while upside targets are first $130, then $146.
My take: Solana has always been talked about as “fast,” but what has held back payments and chain games hasn’t been peak TPS—it’s been “how long until it’s irreversible.” 12.8 seconds means the payee has to keep waiting for over ten seconds with bated breath; 150 milliseconds is what it feels like when you can swipe at a POS machine and just walk away. That’s why this change is worth paying attention to more than a typical version update.
But two things must be kept separate. First, devnet is not mainnet: Anza has explicitly said it hasn’t published a mainnet launch date for Alpenglow yet. Even after the tests run successfully, the switch across the whole network still depends on validator upgrades and stability observation in between. Second, today’s SOL rise isn’t because of this—it's up because of the price story, not the technology story. Don’t stitch the two narratives together.
Do you think compressing finality from 12.8 seconds to 0.15 seconds will let Solana truly pull ahead in the payments track—or is it just another case of “tests look great, mainnet will have to wait”? Let’s discuss in the comments.
Click the avatar to watch the live stream
Every day I’ll keep you on top of the performance of public chains and Solana’s hot topics—not just what’s happening in the news, but also the logic and opportunities behind it 👀🚀
📢 群里有实时分析
On September 26, Solana co-founder Anatoly Yakovenko reposted the news about Alpenglow landing on devnet, with just one line: “You're not ready for this.” Anza’s engineers activated this new consensus at epoch 1167, followed by a countdown—then block production was taken over by the new version. Previously, it had already switched successfully on testnet—which means Alpenglow is now running simultaneously on two public test environments.
What it replaces is Solana’s TowerBFT, which it has used for years. The old logic had validators write votes to the chain—voting while accumulating consensus. Votor, the new approach, moves vote coordination off-chain, then compresses the result into a cryptographic certificate to send back. Votes no longer crowd out block space; the saved capacity is given back to user transactions, and validator operators’ voting costs drop as well. The fault-tolerance assumptions are: up to 20% of staked funds behaving maliciously, and another 20% going offline at the same time—the network still needs to keep moving.
What’s really being rewritten, though, is the “waiting feeling.” Solana now takes about 12.8 seconds for a transaction to reach finality. Alpenglow’s goal is around 150 milliseconds; when validator responses are fast enough, the fast path can be pushed down to about 100 milliseconds. At 150 milliseconds, that’s roughly an 85× improvement.
Put it alongside the price: SOL is currently around $121.4, up about 8.6% over the week. It has reclaimed the $120 level. The market’s watched support range is $115 to $120, while upside targets are first $130, then $146.
My take: Solana has always been talked about as “fast,” but what has held back payments and chain games hasn’t been peak TPS—it’s been “how long until it’s irreversible.” 12.8 seconds means the payee has to keep waiting for over ten seconds with bated breath; 150 milliseconds is what it feels like when you can swipe at a POS machine and just walk away. That’s why this change is worth paying attention to more than a typical version update.
But two things must be kept separate. First, devnet is not mainnet: Anza has explicitly said it hasn’t published a mainnet launch date for Alpenglow yet. Even after the tests run successfully, the switch across the whole network still depends on validator upgrades and stability observation in between. Second, today’s SOL rise isn’t because of this—it's up because of the price story, not the technology story. Don’t stitch the two narratives together.
Do you think compressing finality from 12.8 seconds to 0.15 seconds will let Solana truly pull ahead in the payments track—or is it just another case of “tests look great, mainnet will have to wait”? Let’s discuss in the comments.
Click the avatar to watch the live stream
Every day I’ll keep you on top of the performance of public chains and Solana’s hot topics—not just what’s happening in the news, but also the logic and opportunities behind it 👀🚀
