Anza Activates Solana (SOL) Alpenglow on Devnet With 150ms Finality Target

Devnet Flips to Alpenglow

Solana’s developer network has completed its migration to the Alpenglow consensus protocol, and co-founder Anatoly Yakovenko marked the milestone with a pointed message on X: “You’re not ready for this.” The switch, executed by Anza’s engineering team, puts devnet on a consensus design that cuts transaction finality from roughly 12.8 seconds to 100–150 milliseconds — finality being the moment a payment can no longer be reversed. Alpenglow retires the long-standing TowerBFT algorithm in favor of Votor, a voting system built with researchers from ETH Zurich and signed off by validators through the SIMD-0326 governance process. Its central change moves validator coordination off-chain: technical vote transactions, which had consumed as much as 75% of block capacity, are now folded into compact BLS cryptographic certificates, freeing that hidden bandwidth for user transactions and lowering costs for node operators, whose hardware no longer has to carry vote traffic through blockspace. The Solana Virtual Machine and the smart-contract layer covered in our what is Solana explainer remain unchanged. Anza’s activation sequence — branded Alpenswitch — fired at epoch 1167 after a 17-minute countdown, one day after the same deployment landed on testnet. The new consensus runs a “20+20” fault-tolerance model, staying stable when 20% of validators drop offline and a further 20% of nodes come under attack. When validators holding at least 80% of the staking weight respond, finality lands near 100 ms; the fallback path takes roughly 150 ms. The September rollout across public clusters opens a testing window for dApps where response speed is decisive — payments and Web3 games chief among them — the kind of throughput engineering underpinning Solana’s lead among altcoin platforms.

Xhttps://x.com/toly/status/2103582089673331008?s=20

Validator Votes Leave Blocks

Anza confirmed the devnet switch on Sept. 25, a day after the separate testnet — used chiefly to stress-test validator operations and client software under load — completed its own transition, leaving Alpenglow live on both public clusters. The two environments serve different purposes: devnet lets application teams check software against tokens with no real value, while testnet pushes the network stack to its limits. Alpenglow changes how validators agree on the ledger. Instead of recording their votes as transactions inside blocks, validators exchange those votes directly and reach agreement in one or two rounds. The removal has a counterintuitive side effect: headline transaction totals will fall, because counts that include validator votes shrink even when users make exactly as many payments and trades. The Foundation has asked data providers to recalibrate their historical comparisons so charts are not misread as an activity collapse. Services that assemble transaction histories face a sharper requirement — segregating rival candidate blocks until final selection, because merging their contents could corrupt the record. Applications that merely send transactions and read balances need no migration at all. The settlement economics explain the push: an exchange could release a customer’s deposit sooner, and a payment application — the same rails that self-custodial spending products like Tangem Pay build on — could tell a merchant a sale is final in a fraction of a second. Two caveats temper the timeline. The 150 ms figure remains a target derived from simulations, not yet a demonstrated result under live-market conditions, and wallet front-ends such as the Tangem Mobile Wallet plus exchanges’ own deposit checks add further waiting time. No firm launch date exists for mainnet-beta; Anza’s software schedule tentatively lets routine feature activations restart from Sept. 28, without naming that date as Alpenglow’s mainnet debut. Readers tracking the market in real time can follow live spot and futures prices on MEXC.

confirmed the devnet switchhttps://x.com/anza_xyz/status/2103575642080006615

Mainnet Timing Still Open

Our reading of the rollout: Solana is not tuning its consensus overhead, it is deleting it. SIMD-0326, the proposal validators approved, does not adjust TowerBFT — it retires the mechanism entirely, moving vote messages into BLS certificates and, from devnet epoch 1167, freeing node operators from relaying and storing vote traffic in blockspace. That capacity is the real prize, with vote transactions having consumed up to 75% of block space. As our earlier Alpenglow testnet rollout coverage tracked, the open variable is mainnet-beta: undated while stability monitoring runs, leaving 150 ms finality a target rather than a live-market result. Our September staking inflows report shows the validator stake base still widening — the pool that 80% threshold draws from — and readers weighing exposure can start with our guide to buying Solana.