Coinbase, one of the world’s largest cryptocurrency exchanges, experienced a major service outage lasting more than two hours on May 8, 2026, directly linked to an underlying disruption at Amazon Web Services (AWS). The outage originated in AWS’s US-EAST-1 region in Northern Virginia, in a single availability zone (use1-az4), after a thermal event caused a power loss that damaged hardware.
The disruption hit the same day Coinbase reported Q1 2026 earnings — including a $394 million GAAP net loss and a 14% workforce reduction tied to its AI-native operational restructuring.
This marks the second major AWS-driven Coinbase outage since the October 2025 incident.
A Thermal Event in Northern Virginia, a Two-Hour Outage in Crypto
At approximately 5:25 PM PDT on Wednesday, May 7, 2026, AWS engineers detected problems in the use1-az4 availability zone of the cloud provider’s US-EAST-1 region — a Northern Virginia facility that is both AWS’s oldest and largest hub, and the single most important piece of internet infrastructure in the United States.
The cause, per AWS’s incident reports, was a thermal event — overheating in one of the data center’s halls — that triggered a power loss damaging some EC2 instances and EBS volumes in the affected hardware racks. By 6:47 PM PDT, AWS warned that “other AWS services that depend on the affected EC2 instances and EBS volumes in this Availability Zone may also experience impairments.” AWS engineers were still working to recover racks “in a controlled and safe manner” more than 12 hours after the incident began.
For Coinbase users, the disruption surfaced at 18:53 PDT with the first status-page update: “We are aware that customers may be unable to transact on Exchange at this time. Customer funds are safe.” At 19:38 PDT, the company posted a second update explicitly tying the issue to the AWS outage. Beyond the core Exchange disruption, Solana and ALEO sends and receives also faced delays, while fiat services remained operational.
The outage extended past two hours. At time of publication, AWS recovery efforts remain ongoing.
The Same-Day Earnings Coincidence
The timing is unusually pointed.
After the market close on Wednesday, Coinbase had reported its Q1 2026 earnings — a quarter that produced an all-time high in crypto trading market share (8.6%), retail derivatives revenue past $200 million annualized, and prediction markets scaling to $100 million annualized in just two months. But it also produced a GAAP net loss of $394.1 million, the company’s second consecutive quarterly loss, and a 67% year-over-year drop in Adjusted EBITDA as macro headwinds compressed trading volumes.
The most consequential operational announcement on Wednesday’s call, however, was structural rather than financial. Per the Q1 earnings deck, Coinbase announced a 14% headcount reduction, a $50–60 million one-time restructuring charge, approximately $500 million in annualized cost reduction, a reorganization of engineering teams from a legacy model, and a reported 78% year-over-year increase in pull requests per engineer under the new structure.
CEO Brian Armstrong framed the restructuring on the call as a strategic bet on agent-driven productivity rather than a cyclical response to market conditions. Coinbase, he said, is “at the center of the agent economy” — citing the fact that Base, Coinbase’s Layer-2 blockchain, processes more than 90% of onchain agentic stablecoin transaction volume globally.
Then, hours later, the same blockchain infrastructure that Coinbase positions as agent-economy rails went offline because a single AWS data center in Virginia got too hot.
The contrast is, as one industry observer put it on X late Wednesday, “the entire centralized-cloud-vs-decentralized-finance argument compressed into a single news day.”
A Repeating Pattern: Three AWS Outages in 18 Months
For Coinbase, this is not a new failure mode.
October 2025: A DNS resolution failure at AWS’s DynamoDB API endpoint in US-EAST-1 cascaded across 58 AWS services globally and took down Coinbase, Robinhood, Snapchat, Reddit, Roblox, Fortnite, Ring, Venmo, Hinge, EA, Hulu, Max, and Xbox Network simultaneously. The outage prompted a public statement from Coinbase confirming users were “unable to access” the platform “due to an AWS outage” — language nearly identical to Wednesday’s incident.
Multiple smaller US-EAST-1 incidents: Per the Statusfield outage tracker for Coinbase, the platform has experienced at least eight discrete degraded-performance events in the past month alone — including incidents on April 22, April 23, April 24, April 29, and the May 7–8 outage now in progress.
2021: An earlier major US-EAST-1 outage took large portions of the internet offline.
The structural reason all three traced to the same region is that US-EAST-1 is the default. Northern Virginia hosts 663 active data centers, more than any other state in the U.S., and AWS’s hub there is older, larger, and more deeply integrated into the global internet’s routing tables than any of its other regions. When something goes wrong there, the blast radius extends across consumer services, financial platforms, government systems, and AI applications simultaneously.
The “Decentralized Finance, Centralized Pipes” Problem
The outage reignites a debate the crypto industry has had, without ever fully resolving, for years.
Coinbase markets itself as the rails of the agent economy, the largest regulated stablecoin platform, the operational backbone for 62% of global onchain stablecoin transaction volume on Base, and the storage layer for 12% of all global crypto assets ($294 billion in customer assets on platform as of Q1, per the company’s earnings deck).
Yet when AWS US-EAST-1 has a thermal event in a single building in Northern Virginia, all of that infrastructure becomes unreachable for two-plus hours.
The competitive picture amplifies the contrast. During the October 2025 outage, Binance, Kraken, and OKX remained online because they run their infrastructure across different cloud providers, multi-region deployments, or in some cases on-premises. Crypto-native infrastructure plays — Filecoin, Akash Network, OORT, and others working on decentralized cloud computing — have argued for years that the convenience of single-region centralized cloud is a structural risk the industry has been unwilling to price in.
The argument is more pointed now. In Q1, Coinbase’s own earnings deck described AI agents as the next frontier of finance, with a forecast of $35 trillion in agent-processed transactions by 2030 (citing McKinsey research). If that thesis is correct, the question is whether the rails carrying those transactions can afford to go offline for two hours when an air conditioning unit fails in Loudoun County.
What Was and Wasn’t Affected
Per Coinbase’s status page and public posts, the outage affected:
- Coinbase Exchange transactions (degraded performance, intermittent unavailability)
- Solana network sends and receives (delayed)
- ALEO sends and receives (delayed)
Services that remained operational:
- Buys, sells, and fiat deposits/withdrawals (per the company’s standard outage messaging)
- Customer custody (funds remained secure throughout)
- Coinbase Wallet self-custody (not directly tied to Exchange infrastructure)
- Base Layer-2 chain (operates independently of AWS, though some Base-app integrations may have been affected)
The company has not published a final timestamp for full service recovery and has directed users to the AWS service health page for incident-level updates.
The Q2 Question
For Coinbase, the operational disruption is recoverable. Two hours of degraded Exchange performance during a relatively quiet trading session — the macro picture has been softer through Q1 and into early Q2, with transaction revenue at approximately $215 million quarter-to-date through May 5, per the company’s own outlook — is unlikely to materially affect Q2 financials.
The strategic question is harder. Coinbase has positioned its 14% headcount cut as a structural shift toward an AI-native operating model in which fewer engineers, supported by AI agents, can run a more efficient platform. That model assumes the underlying cloud infrastructure those agents depend on is reliable, multi-region, and resilient enough to support high-stakes financial transactions 24/7.
The outage suggests the assumption may need to be re-examined — particularly as Coinbase’s Bitcoin custody footprint grows and as the company’s role as a stablecoin issuer, derivatives venue, and prediction-market operator places increasingly diverse customer flows on the same operational stack.
The next AWS outage is not a question of if. The question is whether the next one happens during a calm trading session like Wednesday’s — or during a period of high volatility, when a two-hour Exchange downtime translates into far more than reputational risk.
Coinbase’s Update at 09:45 UTC
Coinbase Support took to X to share an update regarding the outage, attributing the disruption to AWS failures in US-EAST-1’s use1-az4 availability zone, where elevated temperatures impaired multiple managed services and EC2 instances.
The post also stated, “This primary issue is now fully resolved – thank you for your patience. If you have any outstanding questions about your account, please reach out to Coinbase Support, we’re ready to help.”
With services now restored, Coinbase plans a full post-mortem using AWS’s upcoming retrospective and directs users with account questions to contact support directly.
Coinbase Resumes Trading After Nearly Seven-Hour Outage
Coinbase has fully restored exchange operations after a disruption that lasted nearly seven hours, with the company confirming that trading systems are back online and stable.
In its latest update, Coinbase said the primary issue tied to the AWS infrastructure failure in the US-EAST-1 region has been resolved, while internal teams continue monitoring platform performance and reconciling delayed activity generated during the outage window.
The exchange also directed users experiencing lingering account or transaction issues to contact customer support as post-recovery checks continue. The incident is likely to intensify industry discussions around infrastructure redundancy and the risks of relying heavily on single-region cloud environments for always-on crypto trading systems.
