TheBlock
Coinbase re-enabled trading on its exchange on Friday after several hours of disruption that pushed every market on the venue through Cancel Only and auction modes before phased restoration. The exchange first flagged the incident on X as "degraded performance," then attributed it to a "broader AWS outage" — later specifying "increased temperatures" inside AWS Availability Zone use1-az4 in the US-EAST-1 region, a single-AZ fault that propagated through Coinbase's matching and order-entry stack.
Solana and ALEO transfers saw delayed sends and receives beginning at 21:00 ET on May 7, with buys, sells, and fiat rails for those networks still working. As the incident escalated, Coinbase told customers they "may be unable to transact on web and mobile," disabled market and limit orders, and shifted all markets to Cancel Only before running an opening auction — limit orders were accepted during the auction but no matching occurred for a minimum of 10 minutes, with crossed orders matched at the opening print once the auction closed.
Why it matters
A venue-wide halt on the largest US-listed crypto exchange is the kind of event traders, market makers, and risk teams re-price around — the question is never whether outages happen, it's whether the blast radius was contained. The single-AZ AWS dependency is the line the rest of the industry reads: the venue's incident ran on US-EAST-1's use1-az4, meaning one hyperscaler region's one availability zone took out order entry, matching, and ancillary network rails on a top-three venue simultaneously. That concentration shows up in every exchange's architecture review after an event like this.
Market impact
Trading is back live and the incident is closed, but the timing amplifies the read: the outage landed days after Coinbase disclosed a 14% workforce reduction as the company restructured around AI-focused systems and tighter cost controls in a weaker market.
Frequently asked questions
-
What caused the Coinbase trading disruption on May 7?
Coinbase attributed the hours-long disruption to "increased temperatures" inside AWS Availability Zone use1-az4 in the US-EAST-1 region, part of a broader AWS outage. The single-AZ fault propagated through Coinbase's order-entry and matching stack.
-
Which Coinbase features were affected and which kept working?
Order matching, market orders, and limit orders were suspended for hours, and Solana and ALEO transfers saw delayed sends and receives from 21:00 ET. Buys, sells, and fiat withdrawals and deposits for those two networks were reported unaffected.
-
How did Coinbase restore trading after the outage?
The exchange moved all markets into Cancel Only mode first, where resting orders could still be cancelled but no new orders were accepted. It then transitioned into an opening auction, accepting limit orders and publishing indicative prices, before resuming matching at the opening print.
-
How did Coinbase stock react to the disruption?
COIN was last seen down 2.53% in pre-market trading on Friday, though the print conflated the outage with a recently disclosed 14% workforce reduction and broader market weakness rather than the operational event alone.
-
Why does an AWS outage take out a crypto exchange?
Centralised exchanges run their matching engines, order databases, and ancillary services on cloud infrastructure, and many concentrate workloads in a single hyperscaler region. When one availability zone degrades — as use1-az4 did on Friday — the venue can lose order entry and matching even if its core systems are…
Crypto News
Benjamin Cowen
Crypto Capital Venture
WatcherGuru
CoinDesk
Crypto Rank News
CoinTelegraph