Alpenglow: What It Means for Institutions, One Week On

Solana's Alpenglow upgrade went live on 21 October, replacing the consensus system the network had run since March 2020. For institutions, one of the biggest changes is settlement. A Solana transfer now becomes irreversible in well under a second, where it used to take several seconds.

Our research report, Solana Alpenglow Upgrade: The Impact on Institutional Staking, covers how the upgrade works. This piece looks at what has changed in practice for the institutions that use Solana: exchanges and custodians, trading firms, staking product issuers and the analysts who track the network.

‍

At a glance

• For exchanges and custodians, a transaction marked "confirmed" is now also final, removing the wait of several seconds before a transfer became irreversible.
• For trading firms, Solana now produces a block every 200 milliseconds and finalises in well under a second.
• For staking products, epochs now last about a day, roughly halving the worst-case time to turn staked SOL back into liquid SOL.
• For analysts, Solana's headline transaction count fell because validator votes have left the chain. User activity held steady.
• Rotor, the second phase of Alpenglow, has no date yet.

‍

Exchanges and custodians: a transfer is final almost as soon as it lands

‍

Before Alpenglow, a Solana transaction had two milestones. It showed as “confirmed” within a second or two, but only became "finalized", or irreversible, 32 slots later. That took about 12.8 seconds at Solana's original slot time, and about 6.4 seconds after slot times were halved in early October [1]. Any institution that treats finality as the moment risk transfers had to wait out that gap or carry the risk.

The Solana Foundation's guidance now states that "confirmed and finalized now describe the same state". Anza, which built Alpenglow, targets finality in roughly 150 milliseconds and names exchanges crediting deposits among the first beneficiaries. Early readings on the Solana Foundation’s own dashboard sit a little above that target, at around 200 milliseconds [2].

The benefit depends on each firm updating its own systems. The Foundation asks integrators to point their code at “finalized” and re-check any timeouts or delays tuned to the old 12.8-second finality. Section 2.8 of our report looks at what faster finality means for custody, redemptions and fund operations.

‍

Trading firms: faster blocks and faster finality

Two changes have landed in quick succession. Between August and early October, Solana cut its slot time from 400 to 200 milliseconds in four steps, so blocks now arrive twice as often [1]. Each validator’s turn as block producer fell from 1.6 seconds to 800 milliseconds. Alpenglow then cut the time to finality.

Solana's core contributors draw a distinction between these two speeds in their Internet Capital Markets roadmap. They argue that how fast a transaction lands matters most for liquidity, because it decides how quickly market makers can update their quotes before stale ones are picked off. They expect landing times to fall towards 50 to 100 milliseconds after Alpenglow [3].

Finality matters most where trading meets settlement, when a position or collateral has to move between venues, custodians or counterparties. Our earlier report, Solana’s Evolution to Institutional Grade Trading Infrastructure, traces how these changes have built up over five years. It is too early to attribute any change in Solana trading activity to Alpenglow itself.

‍

Staking products: shorter epochs, faster liquidity

The slot time cut also halved the length of an epoch, the period at whose boundaries Solana activates and deactivates stake. Epochs have moved from about two days to about one [1]. As a result, the worst-case time to turn staked SOL back into liquid SOL has roughly halved.

This matters most for staked exchange-traded products and for exchanges running pooled customer accounts. Both hold a buffer of unstaked SOL, sized against how long it takes to unstake. When our report was written, one-day epochs were the final, pending stage of the change. That stage went live on 9 October, and section 2.6 of the report sets out how it can let funds stake a higher share of their assets.

Alpenglow also changes the network's resilience assumptions. Solana previously stayed live as long as no more than a third of stake was unresponsive. Under Alpenglow's 20+20 model, it tolerates up to 20% of stake acting maliciously plus a further 20% offline. Our report explains the trade-off and our view of it in sections 2.3 and 3.

Delegators have nothing to do. The Foundation's guidance lists no action for users and describes no change to how rewards reach delegators. Twinstake's validators completed the migration on 21 October, and nothing changes for clients' existing delegations.

‍

Analysts: re-baseline the data

Anyone tracking Solana through transactions per second will have seen a sharp drop on 21 October. Solana Compass data shows that validator votes made up about 62% of processed transactions in the week before the switch [4]. Under Alpenglow, validators exchange votes directly and aggregate them into a single certificate, so those votes no longer appear in blocks.

User activity held steady across the switch. Non-vote transactions ran at about 2,500 per second in the hour of the migration, against about 2,640 in the same hour the previous day [4]. The Solana Foundation calls the fall an accounting change and advises re-baselining dashboards. Any due diligence comparing Solana activity before and after 21 October should strip vote transactions out of the earlier figures.

‍

What comes next

Votor, Alpenglow's new voting system, is the first of its two phases. Rotor, which changes how blocks are distributed across the network, "is not yet scheduled". Anza has also named geographically aware leader rotation and a post-quantum version of Alpenglow as further work, without committed dates.

Our research report covers the full institutional case, including the 20+20 resilience model, the new validator economics and settlement risk. Download Solana Alpenglow Upgrade: The Impact on Institutional Staking.

‍

Disclaimer

Twinstake does not provide staking services to retail customers. This article is not intended as a promotion, offer, invitation or solicitation for the purchase or sale of any investment, nor is it intended to give rise to any other legal relations whatsoever and must not be relied upon for the purposes of any investment decision. It does not constitute financial, legal, or investment advice. If you do not have the relevant professional experience in matters relating to crypto asset investments, you should not consider this article to be directed at you.

References to network performance, rewards, finality and planned protocol upgrades are based on information available as at 28 October 2026, may change and are not guaranteed. Historical reward data is not a reliable indicator of future results. Staking rewards are variable, and staking may involve slashing, downtime, liquidity, market, protocol, operational and other risks. Alpenglow became active on Solana mainnet on 21 October 2026. Performance figures in this article are targets from published proposals or early readings from public sources, and are not guarantees. References to blockchain or protocol finality are technical in nature and do not constitute representations regarding legal settlement finality or the elimination of counterparty, custody, insolvency or other risks.

This report and the information in it are not directed at, or intended to be made available to, retail customers. It is directed only at persons who are professional investors (for the purposes of the Alternative Investment Fund Managers Directive (2011/61/EU) (known as ‘AIFMD’); professional clients or eligible counterparties for the purposes of the Markets in Financial Instruments Directive II (Directive 2014/65/EU) (known as ‘MiFID II’); if you are in the UK, to “Investment Professionals” or “High Net Worth Companies” as defined in Articles 19 and 49 respectively of the Financial Services and Markets Act 2000 (Financial Promotion) Order 2005, or as otherwise defined under applicable local regulations and at whom this report and the information in it may lawfully be directed in any relevant jurisdiction.

The appearance of any third-party hyperlinks or third-party reference in the report does not constitute an endorsement, guarantee, warranty, or recommendation by Twinstake. Do conduct your own due diligence before deciding to use any third-party services.

Twinstake shall have no liability for any loss or damage that may arise directly or indirectly from the use of or reliance on the information provided herein or for any errors or omissions in the information.

‍

‍

Sources

‍

[1] Solana Compass, "Solana 200ms Slot Time Live on Mainnet: SIMD-0525 Complete at Epoch 1053", October 2026. https://solanacompass.com/news/solanas-200ms-slot-time-is-live-on-mainnet-at-epoch-1053-completing-simd-0525

[2] AltcoinBuzz, "Solana Alpenglow Live on Mainnet: 150ms Finality Target, 202ms Observed", 22 October 2026 (reading from the Solana Foundation's Alpenglow dashboard). https://www.altcoinbuzz.io/solana-activates-alpenglow-targeting-150ms-finality

[3] Yakovenko, Resnick, Bruder, Federa, Heaney and Samani, "The Internet Capital Markets Roadmap", Anza, 24 July 2025. https://www.anza.xyz/blog/the-internet-capital-markets-roadmap

[4] Solana Compass, "Solana Alpenglow Live on Mainnet: TowerBFT Retired, Anza's 150ms Finality and What Changes", October 2026. https://solanacompass.com/news/solana-alpenglow-is-live-on-mainnet-as-anza-retires-towerbft-at-slot-459653000

‍