Solana Neared Freeze Threshold as Routing Glitch Knocked Nearly 29% of Staked Tokens Offline, Marinade Finance Says
Key Takeaways
- •A routing failure at Teraswitch's Miami facility temporarily disconnected approximately 29% of Solana's staked tokens, bringing the network close to the one-third threshold that would prevent transaction finalization.
- •Roughly 90 validators across Europe and Asia were affected for approximately 33 minutes and collectively lost 333 SOL in rewards, a sum covered by validator bonds.
- •A single network operator identified as AS2032 controlled more than a quarter of all staked tokens, exceeding Solana's prescribed safety limit and highlighting infrastructure concentration risks.
- •Solana Foundation reported that 597 of 699 staked validators continued operating throughout the incident, characterizing the event as evidence of the network's resilience through infrastructure diversity.
- •The contrasting assessments from Marinade and Solana Foundation raise questions about what metrics beyond raw uptime best capture operational resilience in proof-of-stake networks.

A routing failure at a major data center provider briefly knocked nearly 29% of Solana's staked tokens offline on Wednesday morning, bringing the smart contract blockchain within roughly 20 million tokens of a freeze threshold, according to staking platform Marinade.
Solana stops finalizing transactions — the point at which they become irreversible — if more than one-third of staked tokens go offline. Staking refers to the practice of locking up coins to help secure a blockchain network in exchange for rewards. Marinade said the network came perilously close to that limit, a scenario reminiscent of the five-hour outage Solana experienced in February 2024.
"If delinquency had gone past a third, nothing finalizes for anyone holding SOL anywhere, and there's no bond for that. The February 2024 halt took about five hours to restart," Marinade said in an explainer post on X.
Routing Failure Traced to Teraswitch's Miami Facility
The disruption originated from a bad internet route at Teraswitch's Miami facility and then propagated to data centers across Europe and Asia, severing connectivity for validators in London, Amsterdam, Frankfurt, Singapore, and Tokyo. North American validators remained online throughout the incident.
Teraswitch resolved the issue in approximately 10 minutes, with traffic flowing normally again by 4:16 a.m. UTC.
Roughly 90 validators were affected and collectively lost 333 SOL in rewards, a relatively modest sum that will be covered by "validator bonds." Most of the impacted validators — including the large operator Helius — remained offline for the full 33-minute duration because their backup systems failed to activate. Validator bonds, which some Solana staking protocols require as a form of collateral, are designed to offset exactly this type of penalty — though they do not protect against the broader systemic risk of a near-freeze event.
Concentration Risk Exposed
A single network operator, identified as AS2032, controlled more than a quarter of all tokens locked to secure the network, exceeding the Solana-prescribed safety limit. Nearly all of those tokens went dark simultaneously. Other operators lost an additional 14 million tokens during the same brief window.
The incident underscored the systemic risk of concentrating staked tokens behind a single connectivity provider. If more than one-third of the network's tokens were to go offline at once, the entire blockchain would freeze for every SOL holder, with no rapid mechanism to repair the broader consequences. The episode echoes a wider concern across proof-of-stake networks, where infrastructure and staking concentration has drawn scrutiny from researchers and developers on chains including Ethereum, where dominant liquid staking providers like Lido have prompted debates about whether caps or incentives are needed to preserve decentralization.
Solana Foundation: Network Remained Operational
Solana Foundation's VP of Tech Jacob Creech framed the episode as evidence that the network's design withstood the pressure.
In a post on X shortly after the incident, Creech wrote that an infrastructure provider used by some Solana validators had failed, yet "You probably didn't notice, because the network didn't: blocks kept producing and transactions kept landing."
Creech laid out the key data points: the Solana network remained operational throughout, 597 of 699 staked validators continued voting (roughly six out of seven), the affected validators recovered within 40 minutes, and participants in the Solana Foundation Delegation Program were entirely unaffected.
"Because Solana validators are distributed across independent infrastructure providers, the failure of a single provider did not interrupt the network," Creech said. "This is decentralization and infrastructure diversity working as intended, and a proof point for Solana's resiliency through periods of stress."
The contrasting interpretations from Marinade and the Solana Foundation highlight an ongoing tension in blockchain operations: networks can continue producing blocks while still approaching critical safety thresholds, raising the question of what metrics — beyond raw uptime — best capture operational resilience for users and developers.
Context
Solana is one of the leading smart contract blockchains, with approximately $4.3 billion worth of assets locked in decentralized finance protocols on the network. The blockchain has established itself as a faster and lower-cost alternative to Ethereum, though it has experienced several outages in the past.
Aug. 13, 4:01 UTC: Adds comments from Solana Foundation.