Solana Activates Alpenglow, Its Most Ambitious Consensus Upgrade

Highlights

Solana activated Alpenglow on mainnet-beta at slot 459,653,000, retiring TowerBFT and reducing the time to transaction finality from about 12.8 seconds to roughly 150 milliseconds. The change combines confirmation and finality into a single process, a shift intended to help exchanges, banks, and merchants know sooner when an asset transfer cannot be reversed. The protocol was developed by Anza, tested through multiple stages, and activated by independent validators. Further proposals include Rotor, geographically aware leader rotation, and Quantumglow, but none has a committed date. The upgrade’s speed gains come alongside continuing debate over resilience and the risk of greater validator centralization.

Sentiment Analysis

• Overall tone: The article is primarily factual and measured. It describes a major technical launch and reports concrete improvements without presenting them as guaranteed benefits for every user.
• Positive elements: Alpenglow is reported to reduce finality from about 12.8 seconds to roughly 150 milliseconds. The article also notes that the protocol passed public testing and was activated by independent validators, while describing mechanisms designed to tolerate malicious or offline stake.
• Neutral elements: Technical details about voting, validator communication, mainnet activation, and the upgrade’s development history are presented as explanations rather than endorsements. Future proposals are clearly distinguished from features already live.
• Areas of concern: The discussion of a potential increase to the block compute-unit cap includes criticism that the change could benefit well-funded operators and increase centralization. Quantumglow and other planned work also remain undated.
• Sentiment balance: The coverage is cautiously positive about speed and deployment, while retaining a neutral-to-concerned view of operational trade-offs and future governance decisions.

Article Text

Solana has activated Alpenglow, a new consensus protocol that replaces TowerBFT on its mainnet-beta network. The change went live at slot 459,653,000, at around 14:00 UTC, according to core engineering team Anza. Validators moved to the new protocol on the running network while preserving the existing ledger, coordinating the transition around a shared handoff point.

The upgrade changes how long it takes for a transaction to become final. Finality is the stage at which a transaction can no longer be reversed. Exchanges, banks, and merchants may wait for this point before treating an asset as transferred. Under TowerBFT, a transaction could receive conditional confirmation within seconds, but full finality took about 12.8 seconds. Anza says Alpenglow reduces that interval to roughly 150 milliseconds by combining confirmation and finality into one process.

Anza CEO Brennan Watt described the project as the largest and most ambitious blockchain upgrade since Ethereum’s 2022 switch to proof of stake, commonly called “The Merge.” He said the team took Alpenglow from a white paper to mainnet in under 18 months. The protocol’s design was published in May 2025 by Anza’s research team, led by ETH Zurich professor Roger Wattenhofer alongside Quentin Kniep and Kobi Sliwinski.

Alpenglow changes how validators communicate their agreement. Rather than recording votes in later blocks, validators send votes directly to one another, with each vote weighted by stake. Wattenhofer explained that when 80% of votes support a block, the network can immediately treat that block as final. A second method uses two voting rounds, each requiring 60% of stake. Wattenhofer said this route may be faster when validators are located close to one another, such as within Europe.

The protocol is designed to continue operating when less than 20% of stake behaves maliciously and a further 20% is offline. It also includes a response to prolonged delays: if the network goes 10 seconds without finalizing a block, validators gradually lengthen block times to give messages more opportunity to get through. Wattenhofer said this mechanism is intended to let the network eventually produce a block even under poor internet conditions.

Before mainnet activation, Alpenglow underwent public testing that included red teaming, penetration testing, repeated migration rehearsals, and a security competition. It was then launched on Solana’s testnet and devnet. Anza implemented the protocol in Agave, the validator software it maintains, while independent validators carried out activation. Solana validators approved the upgrade in September 2025, with 98% voting in favor of SIMD-0236, the proposal that implements Alpenglow.

The rollout had been signaled in stages. Alpenglow went live on Solana’s community test cluster in May 2026. At that time, Anza lead economist Max Resnick said the team was targeting mainnet activation in late Q3 or early Q4. Anza confirmed the October 21 date on October 13. Resnick had previously told Decrypt that the most immediate user-visible difference would be confirmation latency of approximately 150-200ms.

Alpenglow may also make other network changes possible, though those proposals are distinct from the live protocol. Rotor is intended to change how block data spreads through the network, while geographically aware leader rotation aims to reduce delays as block production moves between validators. Quantumglow is a research design that would replace Alpenglow’s elliptic-curve cryptography with hash-based signatures. Its goal is to provide security against quantum computers while preserving the same finality speed; it has no committed date.

Another proposal concerns Solana’s block capacity. Developers from Jump Crypto’s Firedancer team have suggested removing the cap of 60 million compute units per block once Alpenglow is live. Under that approach, blocks could grow to the capacity of the fastest validators, while slower validators could use a “skip vote” feature to abstain. Critics have cautioned that the change could advantage well-funded operators and increase centralization. The proposal illustrates that faster finality is only one part of the upgrade’s broader implications: decisions about capacity and validator participation may also shape the network.

Key Insights Table