Coinbase’s latest outage was triggered by an AWS datacentre cooling failure, but Brian Armstrong’s response has shifted attention toward a deeper issue facing crypto exchanges: how much resilience can platforms build before it starts affecting trading speed?
For an industry built around the promise of uninterrupted, always-on trading, even a short outage can become a revealing moment.
Coinbase’s disruption on 8 May did not stem from a cyberattack, market panic or software deployment gone wrong. Instead, according to CEO Brian Armstrong, the issue began with “a room overheating in an AWS datacenter when multiple chillers failed”.
The incident caused degraded performance across parts of Coinbase’s infrastructure, including Advanced Trade API services and delayed Ethereum transaction status updates. While the company stated that customer funds remained safe throughout, the outage quickly reopened a familiar debate about operational resilience, cloud concentration risk, and whether crypto exchanges have prioritised low-latency trading over redundancy.
Armstrong’s public explanation also offered a rare glimpse into the engineering trade-offs underpinning major crypto exchanges.
“We design our services to be redundant to downtime in any one AWS Availability Zone (AZ), and most of our systems worked this way last night, but not all,” he said.
“Our centralised exchange did not. Exchanges have unique architectures that optimize for latency and co-location of clients.”
AWS failure triggers exchange disruption
According to Coinbase’s public status page, users first experienced delayed Ethereum receives and send status updates during the morning of 8 May, while later in the day the company reported degraded performance affecting its Advanced Trade API.
Coinbase state buys, sells, and fiat withdrawals and deposits remained operational during part of the disruption, though some trading functionality was affected. The company later confirmed the incidents had been resolved.
Armstrong said Coinbase would now revisit how its exchange infrastructure handles AWS Availability Zone failures after the outage exposed weaknesses in its failover design.
“It is possible to make exchanges resistant to AZ failures, but this can introduce latency delays that are not desirable along with breaking customer co-location,” he explained.
“Given this incident, we’ll revisit these tradeoffs to ensure we’re giving you the best possible venue to trade.”
The reference to co-location is worth considering, as many high-frequency trading firms rely on physically proximate infrastructure to minimise execution delays. In traditional financial markets, exchanges often balance resilience against the ultra-low latency demands of institutional traders. Armstrong’s comments suggest Coinbase has faced similar architectural decisions.
Cloud concentration risk returns to the spotlight
The outage has also revived wider concerns around the growing dependence of financial infrastructure providers on a small number of cloud vendors.
Over recent years, regulators including the Bank of England, European Central Bank and US authorities have repeatedly examined concentration risk tied to hyperscale cloud providers such as AWS, Microsoft Azure and Google Cloud.
While Coinbase described the root cause as an AWS datacentre cooling failure rather than a broader cloud outage, critics argued the disruption still demonstrated how dependent large crypto platforms remain on centralised infrastructure providers.
Gergely Orosz, author of The Pragmatic Engineer, wrote on social media: “This outage is because Coinbase seems to have a hard dependency on AWS, and when AWS (or a part of it) is down, so is Coinbase.”
He added that the incident came “a few days after their CEO said how non-technical teams are shipping code to production”.
Others questioned whether Coinbase’s historical infrastructure spending decisions may have contributed to operational vulnerabilities, though no evidence has emerged directly linking the outage to cost reductions.
Social media commentator Rho Rider claimed Coinbase had reduced infrastructure and developer spending in recent years while prioritising profitability. Coinbase has not publicly responded to those allegations.
The company’s 2023 financial filings did reference lower infrastructure costs tied to “modernization and efficiency” initiatives, though the filings did not suggest these changes compromised platform resilience.
Crypto exchanges face familiar financial infrastructure pressures
For crypto firms increasingly positioning themselves as institutional-grade infrastructure providers, operational resilience has become more than a technical issue.
Coinbase has spent the past several years expanding beyond retail crypto trading into payments, stablecoin infrastructure, custody and institutional services. The company is also one of the largest beneficiaries of growing stablecoin adoption through its relationship with Circle and USDC.
This evolution means outages are now likely to face scrutiny not only from traders, but from payment firms, institutional clients and regulators assessing whether crypto infrastructure can support broader financial activity.
The incident also arrives as crypto firms continue pushing deeper into areas traditionally associated with mainstream financial infrastructure, including cross-border settlement, tokenised payments and real-time transaction processing.
Armstrong said Coinbase would release a detailed technical summary of the incident once completed.
“Thank you to the AWS and Coinbase teams for working through the night to mitigate the issue,” he added.
