Coinbase hits 7-hour outage

Crypto exchanges sell speed and constant availability. That is the whole pitch — markets never sleep, and the pipes are supposed to stay open. But late on May 7, Coinbase got knocked sideways for hours when an Amazon Web Services failure in Northern Virginia spilled into the exchange’s core systems. By early May 8, trading was back. The awkward part is that this landed basically right on top of an already rough earnings week. ### What actually broke? The trigger was not a crypto-specific bug. AWS had an overheating problem in a single US-EAST-1 availability zone, which disrupted EC2 instances and EBS volumes. That matters because those are the boring but essential building blocks underneath a lot of internet services. Coinbase said the exchange’s disruption was tied to increased temperatures in the affected AWS service. In other words — a physical data-center problem bubbled all the way up into one of the biggest crypto venues. (status.coinbase.com) ### How long was Coinbase down? Coinbase’s own status history is the cleanest timeline. It lists “Degraded Performance” from 5:56 p.m. PDT on May 7 to 2:28 a.m. PDT on May 8. Before full reopening, Coinbase said it would first put markets into “Cancel Only” mode, which is the exchange equivalent of cracking the door open carefully before letting normal trading resume. So the headline “7-hour outage” is directionally right, even if the exact customer experience probably varied by product and moment. (finance.yahoo.com) ### Were funds at risk? Coinbase’s message through the incident was that customer funds were safe. That is an important distinction. An outage like this can stop access, delay orders, and create ugly execution gaps, but it is not the same thing as a custody failure or missing assets. For traders, though, being unable to enter or exit during a volatile market can still be expensive. “Safe” does not mean “harmless.” (status.coinbase.com) ### Why does AWS concentration matter here? Because centralized exchanges are only as continuous as the infrastructure stack underneath them. If one cloud region, one availability zone, or one tightly coupled dependency fails, the exchange can end up looking a lot less like a 24/7 market and a lot more like a normal website with a very bad day. This is the single-point-of-failure argument crypto critics and self-custody advocates keep making — not that Coinbase is uniquely fragile, but that centralization quietly reintroduces old infrastructure risks. (status.coinbase.com) ### Why did the timing look so bad? Because Coinbase had just reported a weak first quarter on May 7. Search results pulling from the post-earnings analyst notes show revenue at $1.41 billion, below the $1.48 billion consensus, with adjusted EBITDA of $303.3 million. Barclays cut its price target to $107 from $140 and kept an underweight call. So investors were already focused on softer trading activity, and then the platform itself spent hours impaired. (finance.yahoo.com) That is a rough combo. ### Did the stock collapse anyway? Not really. Coinbase’s investor site shows COIN closing at $201.16 on May 8, up from $192.96 on May 7. That does not mean the outage was shrugged off. It means the stock was trading inside a bigger mix of forces — crypto prices, earnings digestion, and whatever investors think matters more long term: one ugly incident or Coinbase’s broader role in the market. (msn.com) ### So what’s the real takeaway? This was not just a bad night for one app. It was a reminder that “always-on” finance still depends on very physical things — cooling systems, data centers, and cloud architecture choices. Coinbase restored trading. AWS largely resolved the underlying issue. But the episode gave traders a vivid demo of the catch with centralized crypto rails: when the stack underneath them fails, the market can disappear right when people most want access. (investor.coinbase.com) (msn.com)