On April 20, 2026, a series of significant technical failures resulted in the near-simultaneous global outage of the world’s leading artificial intelligence platforms. OpenAI, Anthropic, Microsoft, and Google all reported substantial service disruptions beginning at approximately 08:15 UTC. The event, which lasted for several hours, represents one of the most widespread infrastructure failures in the history of the generative AI sector, affecting both consumer-facing applications and enterprise-grade API services.
OpenAI was the first to acknowledge the instability, updating its official status page to reflect a Major Outage across ChatGPT and the Codex API. According to the company’s technical logs, the issue primarily affected the GPT-4o and GPT-5 preview models, with users encountering Internal Server Error messages and persistent timeouts. At the height of the disruption, third-party monitoring service DownDetector recorded a peak of 165,000 user reports for ChatGPT within a single hour, indicating a global impact spanning North America, Europe, and Asia.
Simultaneously, Anthropic confirmed that its Claude.ai platform and associated API tiers, including Claude 3.5 Sonnet and the recently released Claude 4.0, were experiencing 504 Gateway Timeout errors. Microsoft Copilot, which relies heavily on OpenAI’s underlying architecture, saw widespread failures across its integration points in Windows 11, Microsoft 365, and the Bing search interface. Users reported that while the basic search functionality remained operational, the AI-driven Compose and Chat features were entirely unresponsive.
Google Gemini also faced significant downtime, though the company categorized the event as a partial service disruption. While the Gemini mobile application remained accessible for a subset of users, the Gemini Advanced web interface and the Vertex AI platform used by developers were largely offline. Google’s Cloud Status Dashboard cited unforeseen issues with regional load balancing as a contributing factor, though it did not provide a detailed post-mortem during the event.
By 10:45 UTC, OpenAI’s engineering team issued a statement confirming they had identified a networking configuration issue that impacted traffic routing to their primary compute clusters. Anthropic and Microsoft issued similar updates shortly thereafter, suggesting a potential shared dependency on a major Tier 1 content delivery network or cloud infrastructure provider. Service restoration began gradually at 12:30 UTC, with most platforms returning to full operational capacity by 13:15 UTC. Total downtime for the most affected services was recorded at approximately five hours, impacting an estimated 200 million users worldwide.