Solana Came Within 4.5% of Losing Finality After One Hosting Provider’s Routing Error

Solana came dangerously close to temporarily losing its ability to finalize transactions after a networking error at a single infrastructure provider knocked 28.83% of all staked SOL offline. Solana stops achieving finality when more than 33.34% of stake becomes unavailable, meaning the August 12 incident pushed the network roughly 86% of the way toward that critical threshold before engineers restored connectivity. 

The incident did not result from a hack, blockchain exploit, validator software bug, or attack against Solana itself. Instead, infrastructure provider Teraswitch traced the disruption to a malformed default network route that caused a large number of Solana validators hosted through its infrastructure to suddenly lose connectivity. The blockchain continued operating and ultimately maintained finality, but the incident exposed how reliance on shared internet infrastructure can create unexpected concentration risks for decentralized networks. 

Nearly 29% of Solana’s Staked Supply Suddenly Went Offline

The incident began Wednesday morning when validators representing 28.83% of Solana’s total active stake abruptly became unreachable.

That’s an enormous number for a proof-of-stake blockchain.

Solana requires more than two-thirds of active stake to participate in consensus to continue finalizing blocks. If unavailable stake had crossed 33.34%, the network would have been unable to finalize new transactions until enough validators returned online. 

With 28.83% already offline, Solana was just 4.51 percentage points away from crossing that threshold.

The network therefore got approximately 86% of the way toward losing finality.

One Incorrect Internet Route Caused the Problem

The failure originated with Teraswitch, an infrastructure and connectivity provider used by a significant number of Solana validators.

Teraswitch normally uses a default route internally to indicate that an edge router can reach the broader internet.

During the incident, one of those routes was incorrectly distributed.

That routing mistake effectively caused large numbers of validators relying on Teraswitch’s infrastructure to lose network connectivity simultaneously. 

The validators themselves hadn’t necessarily failed. They simply could no longer reliably communicate with the rest of the Solana network.

Once Teraswitch identified and corrected the routing issue, validator connectivity recovered.

Solana Never Actually Lost Finality

Despite how close the network came, Solana continued finalizing transactions throughout the incident.

That distinction is important.

The blockchain did not halt, transactions were not reversed, and there is no indication that user funds were compromised.

Instead, the event demonstrated how close the network could come to a consensus disruption when a large concentration of validators shares a common infrastructure dependency. 

The incident is therefore better described as a near miss rather than another Solana outage.

Why the 33.34% Threshold Matters

Solana uses proof-of-stake consensus, meaning validators participate in confirming the blockchain based partly on the amount of SOL delegated to them.

For transactions to reach finality, more than two-thirds of stake must effectively agree on the state of the blockchain.

If more than one-third becomes unavailable, the remaining validators no longer have enough participating stake to satisfy that requirement.

The network can continue producing activity under some circumstances, but it cannot provide the same cryptographic certainty that finalized transactions will not be reversed.

That makes 33.34% unavailable stake a critical boundary.

Solana reached 28.83%.

That is why what might otherwise appear to be a relatively ordinary internet routing problem became a significant blockchain infrastructure event. 

The Bigger Problem Is Infrastructure Concentration

The incident raises an important decentralization question.

A blockchain can have thousands of independent validators while still depending heavily on a relatively small number of hosting companies, data centers, internet providers and networking services.

If too many validators use the same provider, an outage affecting that company can suddenly behave like a much larger blockchain failure.

In this case, a single routing configuration mistake was capable of disconnecting validators controlling almost 29% of Solana’s active stake.

That doesn’t mean Teraswitch controlled those validators. The validators remained independently operated.

But they shared an infrastructure dependency.

That distinction highlights an increasingly important concept for decentralized networks — validator decentralization and infrastructure decentralization aren’t necessarily the same thing.

Solana Has a History of Network Reliability Problems

Solana has faced several significant network disruptions throughout its history, although this incident was fundamentally different from many previous outages.

Earlier Solana disruptions have involved problems such as excessive transaction traffic, consensus issues, validator software problems, and bugs.

This time, the underlying Solana blockchain software wasn’t responsible.

The network was instead affected by infrastructure outside the blockchain itself.

That makes the event particularly interesting because improving Solana’s software alone cannot completely eliminate this category of risk.

Validators also need geographic, network, hosting, and infrastructure diversity.

SOL Price Shows Little Reaction

The incident did not trigger a significant immediate market reaction.

At the time of Decrypt’s report, SOL was trading around $76, with relatively modest daily movement. 

That muted response likely reflects the fact that Solana continued operating and user assets were not compromised.

Had the network actually crossed the one-third threshold and stopped finalizing transactions, the market reaction could have been considerably different.

What This Means for Solana

Solana successfully survived the incident, but the near miss exposed a weakness that matters for every proof-of-stake blockchain.

Decentralization isn’t only about how many validators exist. It also matters where those validators run and which infrastructure providers they depend on.

Thousands of independently operated validators can still develop a hidden single point of failure if too much stake relies on the same hosting company, cloud provider, data center, internet backbone, or networking service.

In Solana’s case, one incorrectly configured route at one provider was enough to remove 28.83% of active stake from consensus and bring the blockchain within just 4.51 percentage points of losing finality. 

Nothing was hacked. No private keys were stolen. No smart contract was exploited. And Solana itself didn’t break.

Yet the network still came remarkably close to a serious consensus disruption.