Solana remains online, and 102 of 699 verifiers stop voting

Solana stayed running despite infrastructure failure, and some verifiers were briefly offline.

Jacob Creech, technical director of the Solana Foundation, said that an infrastructure failure on August 12 briefly affected some verifier networks, but the Solana network remained operating as a whole.

Among the 699 pledge verifiers, 597 continued to participate in voting, and the network generated blocks normally. The 102 validators temporarily stopped voting, and affected operators resumed within 40 minutes of the failure. About 28.83% of pledged SOLs have entered a "lost connection" status, which is close to the threshold of 33.34%. Once this ratio is exceeded, the transaction will finally stop.

Teraswitch said that an incorrectly formatted route caused outages at its 12 sites in Europe and Asia, and engineers subsequently restored the connection. Solana's status page does not record main-network events and shows that the cluster has a normal operating rate of 100% over the past 90 days.

Of the 699 pledge verifiers, 597 continue to vote, and blocks and transactions continue to be processed. Affected validators recovered within 40 minutes. Creech said validators in the Solana Foundation commissioned program were not affected. No mainnet events were recorded on Solana's official status page on August 12 or 13, and the Mainnet Beta cluster has a normal operation rate of 100% in the past 90 days.

Solana verifiers remain final despite significant decline in pledge ratios

Although the main network remains online, independent analysis shows that the final impact of the incident is closer to the tipping point than the sheer number of validators suggests. Marinade Finance found that approximately 28.83% of pledged SOLs were in a "lost contact" state within approximately 33 minutes. Solana requires more than two-thirds of pledge participation to achieve transaction finality, so the relevant offline threshold is 33.34%.

Marinade identified approximately 90 validators affected by routing failures, while Creech gave a 597/699 figure that meant 102 validators did not vote at one time. This difference stems from different statistical standards, rather than all 102 validators experiencing the same infrastructure failure.

Creech described the incident as "a demonstration of Solana's resilience." The network did withstand interruptions, but Marinade's data also showed that the proportion of lost pledges reached about 86% of the level needed for an eventual cessation.

"Last night, some of the infrastructure providers used by Solana validators went down. You may not have noticed it, because the network does not: blocks continue to be generated and transactions continue to be implemented. Fact: -Solana network stays up-597 of 699 pledge verifiers..."--Jacob Creech released on August 12, 2026

Teraswitch routing failure spreads from Miami to Asia

Teraswitch's status report traces infrastructure issues to a malformed default route generated by its MIA1 facility in Miami. A route reflector in Amsterdam propagated the erroneous route to European and Asia-Pacific markets, where local routers preferred the erroneous route over the effective route. Twelve stations in London, Amsterdam, Dublin, Frankfurt, Singapore and Tokyo lost accessibility. North American sites were not affected.

Engineers identified the wrong route within 10 minutes and removed Miami from the private backbone network. Service resumed at 04:16:15 UTC time. Teraswitch then deployed global changes at its compute site to prevent similar erroneous routes from blocking traffic forwarding. The provider said the underlying defects were still under investigation and a full root cause report would be released subsequently.

This incident also exposed the concentration of infrastructure among verifiers. Marinade calculated that an autonomous system holds approximately 118.9 million SOLs, accounting for more than a quarter of all pledged SOLs, of which approximately 94% of pledges are offline at the same time.

Solana avoids a repeat of the 2024 network shutdown

This results in contrast to the February 2024 network restart due to the stop of block production. At that time, the validator needed to coordinate the restart, and Solana was offline for nearly five hours. The infrastructure failure on August 12 did not require a restart of the main network.

Solana has since further enhanced its resilience with independent validator software. Firedance will start generating Solana's main network blocks in 2026, adding another validator client path to the dominant Agave ecosystem.

The latest outage tests another decentralization dimension: physical hosting and network connectivity, rather than verifier software. Solana continues to process transactions, but pledges are concentrated on common infrastructure, resulting in a single provider failure that allows a large number of voting pledges to be removed simultaneously.

Follow-up progress

Teraswitch's instant configuration fix has been deployed, but the investigation has not yet been completed. The provider is still checking why the Miami default route was advertised with the wrong attribute and has contacted its hardware provider. A complete report will be released after the relevant work is completed.

Verifier operators may also face greater scrutiny of infrastructure redundancy. Marinade said it plans to review concentration limits by autonomous systems and data centers and provide more transparency on automated failover arrangements. For Solana, the next test will be whether these infrastructure changes can reduce the proportion of pledges exposed by a single routing failure.