Solana Alpenglow upgrade has reached its public developer network. The upgrade lets application teams test a system. That system aims to cut transaction finality from about 12.8 seconds to roughly 150 milliseconds.
Why the Solana Alpenglow upgrade matters
According to the Solana Foundation, Alpenglow is now active on devnet and testnet. However, it has not been activated on mainnet. Anza develops Solana’s core validator software. Anza announced the devnet switch on September 25. That was one day after testnet completed its transition.
The two networks serve different purposes. Application teams use devnet to check their software. They test with tokens that have no real value. Meanwhile, testnet gives validators and infrastructure operators a place to test. They test under more demanding conditions.
For developers, the devnet stage means they can check applications against Alpenglow. They do not need to wait for the system to reach mainnet. Solana’s mainnet continues to use TowerBFT. Therefore, the 150-millisecond figure remains a target. It is not a finality time available to users today.
How the Solana Alpenglow upgrade changes finality
Under TowerBFT, validators submit votes as transactions. These transactions appear inside blocks. Enough votes must accumulate across 32 slots. A block becomes final after that. Currently, this takes about 12.8 seconds.
Alpenglow’s first phase is called Votor. Under Votor, validators send votes directly to one another. The Foundation says a block can reach finality after one voting round. This happens if validators representing at least 80% of stake vote to accept it. A second round provides another path if the first does not meet that threshold.
Finality is the point at which the network agrees on a transaction. At that point, the transaction can no longer be reversed under consensus rules. A faster result could matter to a U.S. exchange. It could help decide when to credit a Solana deposit. It could also help a payment provider decide when a sale is complete. However, each service may still apply its own checks.
The Foundation separates finality from block production time. In September, Solana reduced its target slot time. It cut the time from 300 milliseconds to 250 milliseconds. A further reduction to 200 milliseconds is planned under a separate upgrade. Shorter slots change how often the network produces blocks. Alpenglow changes how validators agree that a block is final.
Data services face changes before mainnet
For applications that send transactions and read balances, the Foundation says no migration is required. Transaction execution, fees, and formats remain the same under the consensus upgrade.
However, services that build transaction histories have more work. Alpenglow can expose competing candidate blocks for the same slot. This happens before the network selects one. The Foundation tells data providers to keep candidates separate. Then they should retain the block that reaches confirmation. Combining transactions from different candidates could leave an explorer with an incorrect record.
Validator votes will also disappear from blocks. They will no longer be submitted as transactions. As a result, a chart that counts both user transactions and validator votes will show a lower total after activation. This happens even if users make the same number of payments.
Some services also read validator participation from vote transactions. Under Alpenglow, that information moves to certificates attached to block data. Those services must change where they obtain it. Operators using Geyser or gRPC data streams must also account for new identifiers. These identifiers distinguish candidate blocks within a slot.
These changes make devnet testing relevant to exchanges and explorers. It also matters to firms that rely on transaction records. Their deposit rules remain their own operational decision. The network upgrade does not automatically change when a platform makes funds available.
Mainnet activation still has no firm date
Solana’s earlier roadmap tied the mainnet rollout to Agave 4.3. It also set an October target. Neither the testnet transition nor the devnet activation sets a confirmed date. Anza’s schedule tentatively allows mainnet activations to resume on September 28. However, that date is not identified as Alpenglow’s activation date. The Foundation’s status page still lists the upgrade as inactive on mainnet.
The Foundation describes Votor as the first phase of Alpenglow. A later phase, Rotor, is planned to replace the block-spreading system. The current rollout concerns voting and finality changes. The roughly 150-millisecond target comes from testing and simulations. It does not come from transactions settled under live-market conditions.