---

An overheating event at an Amazon Web Services data center in Northern Virginia knocked Coinbase Exchange offline for roughly two hours on May 7 and into May 8, 2026, and disrupted services across six blockchain networks: Polygon, Optimism, Arbitrum, Linea, Base, and Scroll. The failure exposed how deeply a supposedly decentralized industry relies on a single physical region for its day-to-day operations. It originated in Availability Zone use1-az4 of AWS's US-EAST-1 region, where a cooling system breakdown caused power loss to hardware and cascaded outward to exchanges, wallets, and layer-2 networks worldwide.

Coinbase confirmed in a public status update that its degraded performance was a direct consequence of the AWS incident. "This issue is related to a broader AWS outage. We're monitoring closely and working to restore full service," the company wrote. AWS said it had brought additional cooling capacity online and was seeing "early signs of recovery," with plans to restore hardware in a controlled and safe manner. Coinbase confirmed that customer funds were not at risk at any point during the disruption.

The timing was commercially painful for Coinbase. The outage coincided with the company's Q1 2026 earnings release, which showed total revenue of $1.41 billion, a 31 percent year-over-year decline and a miss against analyst consensus of $1.48 billion. The company posted a net loss of $394 million for the quarter and adjusted EBITDA of $303 million. Coinbase also disclosed that it purchased $88 million worth of Bitcoin for its own balance sheet during Q1, a notable self-custody signal that sat in sharp tension with the infrastructure failure playing out in real time. COIN shares fell 2.53 percent on the day. Bitcoin was trading at $79,472.88, down 2.34 percent, during the disruption window. Coinbase did report some brighter metrics, including an all-time high global market share of 8.6 percent (spot and derivatives combined) and derivatives volume growth of 169 percent year-over-year, but the narrative on a day that should have been dominated by earnings was instead shaped by an overheated server room.

The Coinbase outage was only part of the picture. Infura, the RPC infrastructure provider owned by Consensys that serves as the default connection point between MetaMask and the blockchain networks it supports, was also disrupted. When Infura went down, MetaMask users on Polygon, Optimism, Arbitrum, Linea, Base, and Scroll lost effective access simultaneously. Base, the Ethereum layer-2 network operated by Coinbase, acknowledged reduced capacity during the incident. Separately, CME Group reported technical and latency problems with its CME Direct trading platform around the same time, though CME did not confirm any link to the AWS event. CME's statement said only that essential maintenance had been completed and users could log back in.

The infrastructure numbers behind this outage are worth sitting with. According to data from Ethernodes cited by CryptoSlate, AWS hosts approximately 2,368 Ethereum execution layer nodes, representing around 37 percent of the entire Ethereum network. A single regional overheating event is therefore capable of degrading a substantial share of Ethereum's node infrastructure, along with every application layer that depends on it. Lefteris Karapetsas, founder of the portfolio tracking tool Rotki, put it bluntly: "The whole vision behind blockchain was decentralized infrastructure, which we have completely failed on."

For users in Nigeria, Kenya, India, and Pakistan, the failure carries consequences that go beyond a frustrating two-hour delay. Africa has more than 350 million unbanked adults and one of the world's fastest crypto adoption rates, growing at roughly 52 percent annually according to Ripple Insights. In those markets, crypto is frequently the most accessible tool for remittances, inflation hedging, and cross-border payments, not a speculative trade to be reopened tomorrow. When an exchange is down or a MetaMask connection fails, users may face locked remittance flows or forced exposure to local currency depreciation during the access window. Unlike US-based traders who can shift to alternative brokerage platforms, many users in sub-Saharan Africa and South Asia have limited substitute on-ramps. Polygon, the network with particularly strong roots in the Indian developer and retail community, was among those cut off through Infura. Pakistan, consistently ranked among the world's top countries for crypto adoption, has a large and active Web3 developer community that relies heavily on MetaMask, making the Infura disruption especially acute there as well. These users tend to access services during off-peak US hours, but US-EAST-1 geography does not discriminate by time zone.

This is not a one-off event. The October 2025 AWS outage in the same region took down Coinbase, Robinhood, MetaMask, and multiple layer-2 networks, affecting more than 3,500 companies across more than 60 countries. During that incident, Base's block finalization times spiked from 14 minutes to 78 minutes and throughput fell by 40 percent. CME Group suffered its own separate data center cooling failure in November 2025 and again in February 2026. A cooling failure at AWS facilities in the UAE in March 2026 extended the pattern further. The May 2026 incident is the fifth major infrastructure event of this kind in seven months. Projects in the Decentralized Physical Infrastructure Network (DePIN) sector, including Akash Network, Fluence, Filecoin, and Arweave, are drawing renewed attention as potential long-term alternatives, but Ahmad Shadid, CEO of O.XYZ, noted that competing with AWS would require "comparable data center density," involving massive electricity investments and sufficient consumer GPU resources, and that those questions currently have no clear answers. Until they do, a single room overheating in Virginia remains a global crypto risk.