The incident began at approximately 5:25 PM PDT when overheating in availability zone use1-az4 of AWS's us-east-1 region triggered a power loss affecting EC2 compute instances and EBS storage volumes. Coinbase told users that "customer funds remain safe" but confirmed degraded performance across its platform. Robinhood also reported service disruptions. By 8:06 PM PDT, AWS said recovery was "slower than originally anticipated," with only "incremental progress" made toward restoring normal temperatures.

AWS confirmed it shifted traffic away from the affected zone for most services and warned customers of longer-than-usual provisioning times.

A Region With a Record

us-east-1 is the largest concentration of cloud infrastructure on the planet, and it is also the most frequently disrupted. Coinbase is not a first-time casualty. In October 2025, an AWS DNS failure affecting DynamoDB severely disrupted Coinbase services: login was left "unavailable or severely degraded" while withdrawals, deposits, and transfers were "delayed or failed entirely." That incident cascaded across 58 AWS services globally, hitting Binance, KuCoin, Infura, Base (Coinbase's Ethereum layer-2 network), and more than 50 other digital platforms. Before that, an AWS outage in April 2025 affected Binance, KuCoin, and MEXC. In February 2026, a data centre fire at AWS's UAE facility caused a 36-hour outage touching the Middle East, South Asia, and Africa. Then on March 1 and 2, 2026, structural fires at the same UAE facility cascaded into the Bahrain region and sent ripple effects across global services.

Following the October 2025 incident, Coinbase published a public retrospective stating it was "exploring all options, including reviewing their regional deployment strategy to implement immediate and long-term fixes to reduce the impact of these types of outages." The May 7 event is the first significant test of whether those reviews translated into action.

The On-Chain Problem

The disruption is not just a customer service issue. It reflects a structural contradiction at the core of the crypto industry.

Approximately 2,368 Ethereum execution-layer nodes, roughly 37 percent of the global total, run on AWS infrastructure, according to Ethernodes data cited by CryptoSlate. When all cloud providers are counted together, around 70 percent of Ethereum nodes run on commercial cloud services, with three providers alone accounting for more than two-thirds of all Ethereum nodes. A full us-east-1 failure would not halt the Ethereum blockchain itself, but it would severely degrade user-facing access and expose stakers running validators on AWS-hosted nodes to inactivity-leak penalties, a gradual balance drain built into Ethereum's proof-of-stake consensus design that accrues whenever validators go offline.

Infura, the dominant RPC gateway that applications use to read from and write to Ethereum, was disrupted during the October 2025 outage, cutting off user access to Polygon, Optimism, Arbitrum, Linea, Base, and Scroll simultaneously. The October 2025 precedent suggests that if the current incident deepens, that exposure pattern could repeat.

Lefteris Karapetsas, founder of Rotki, captured the tension plainly after the October 2025 event: "The whole vision behind blockchain was decentralized infrastructure, which we have completely failed on."

Beyond North America

Users outside the United States carry specific vulnerabilities here. In India, crypto adoption is heavily mobile and P2P-driven, with over 65 percent of P2P trades executed via mobile apps according to Traders Union, and diaspora users frequently rely on platforms like Coinbase to move funds across borders using USDT or USDC. The May 7 incident struck at approximately 5:55 AM IST on May 8, a quieter trading window for South Asian users, but the underlying risk remains: a daytime outage in that time zone would freeze cross-border payment flows with little warning.

In Africa, Nigerian users predominantly access crypto through Binance P2P, Noones, and Paxful rather than Coinbase, which limits direct impact in this incident. However, Binance itself was hit by prior AWS outages, and users in Kenya, Ghana, and South Africa who access DeFi protocols through Ethereum RPCs (typically via Infura or Alchemy) face the same indirect disruption vector that emerged in October 2025. The February 2026 AWS UAE fire already demonstrated that African fintech infrastructure carries concentrated single-vendor risk that regional regulators, including Nigeria's SEC and Kenya's CMA, have not yet publicly signalled they plan to address within their crypto frameworks.

Compounding a Bad Day for Coinbase

The outage landed on the same evening Coinbase reported a surprise Q1 2026 earnings miss. The exchange posted a loss of $1.49 per share against analyst expectations of a $0.27 profit. Transaction revenue came in at $755.8 million, below the $805.2 million consensus estimate. Subscription and services revenue also fell short, at $583.5 million against an expected $619.3 million. COIN stock fell approximately 4 percent in after-hours trading.

The practical path forward for the industry is well understood if not yet widely implemented: multi-region deployments, fallback RPC configurations pointing to geographically diverse providers, and multi-cloud strategies for exchange operators who currently treat AWS as a single point of dependency. Each new outage adds urgency to those conversations. Whether they produce structural change or another round of retrospectives is the open question.