ChatGPT hit by roughly hour-long worldwide outage as logins and signups fail

ChatGPT suffered a worldwide outage for roughly an hour beginning around 8 p.m. ET on August 19, 2026, blocking logins, new-account signups and chat loading, OpenAI said. Service was fully restored by about 9 p.m. ET.

The disruption reached beyond the chatbot. OpenAI’s Codex coding platform and roughly a dozen of the company’s API endpoints were also listed as affected on its status page, meaning developers building on those services felt the outage alongside consumers.

Some users trying to send messages saw errors reading ‘too many concurrent requests.’ OpenAI acknowledged the problem on its status page, saying it had identified login issues across the affected services and was working on implementing a mitigation.

The outage is a reminder of how much routine work now routes through a single provider. ChatGPT reaches hundreds of millions of weekly users, and an hour of downtime spanning logins, Codex and the API touches consumers and businesses at the same moment.

OpenAI has not published a detailed post-incident breakdown of the root cause. The company’s status page recorded the incident as resolved within about an hour but did not attribute it to a specific system failure in the notice. Brief as it was, the episode fits a pattern of concentrated dependence on a handful of AI providers whose stumbles cascade widely.