Solana came within a hair's breadth of a network halt: a provider outage took nearly 29% of staking offline.
On August 12, the Solana network nearly faced a complete halt in block finalization. The cause was a routing failure at the infrastructure provider TeraSwitch, which simultaneously disconnected validators controlling 28.83% of all staked SOL. This value came close to the critical threshold of 33.34%, beyond which the network loses its ability to finalize transactions.
The incident affected approximately 90 validators. TeraSwitch's autonomous system (AS20326) services 118.9 million SOL, accounting for 27.34% of the total staking volume on the network. At the moment of the failure, 94% of these capacities went offline simultaneously. Fortunately, the critical mark was not surpassed: routing was restored in about 33 minutes, after which the validators returned to operation. However, during the downtime, they collectively missed out on approximately 333 SOL in rewards.
A Symptom of a Systemic Problem
The Marinade Finance team, which recorded the incident, rightly points out that what happened is a vivid example of the risk of infrastructure concentration. Such failures are not a coincidence but a natural consequence of the fact that a significant portion of the network depends on one or several large hosting providers. The project has already announced a review of limits on the distribution of validators across autonomous systems and data centers, as well as plans to strengthen backup infrastructure.
It is important to understand that the problem is not new: as of July 22, TeraSwitch was already servicing validators with 27.1% of staked tokens. It was followed by UAB Cherry Servers (12.7%) and Latitude.sh (11%). This concentration creates a systemic risk that has been discussed for a long time, but only now has it manifested in such an acute form.
Some major operators have already recognized the danger. For example, Coinbase reported in its first-quarter report that it distributes its 13 validators through TeraSwitch, and another 10 through Latitude, with each using a backup server in a different location. This is a sensible approach that minimizes the consequences of a single provider's failure.
It is worth recalling that in November 2022, German hosting provider Hetzner already disconnected servers with Solana nodes, affecting about 40% of validators. The network survived then, but the current case demonstrates that the margin of safety is shrinking. The number of active validators in January 2026 fell to 800—the lowest since 2021—whereas at the peak in 2023, there were more than 2,500. This is an alarming signal: the network's decentralization is weakening, and each such failure only accelerates this process.
My analysis: The TeraSwitch incident is not just a technical failure but a warning for the entire Solana ecosystem. If infrastructure concentration continues to grow, the next failure could be fatal. The network urgently needs to diversify validators across providers; otherwise, we risk seeing not a minute-long pause but a complete network halt with unpredictable consequences for price and trust.