On Thursday, June 12, 2025, at approximately 2 p.m. EDT, a massive internet outage erupted across Google Cloud’s infrastructure as well AWS, OpenAI, Anthropic, and more — leading to widespread disruptions in connectivity and service across the internet.

As a downstream consequence of this outage, all major LLM-based services (ChatGPT, Anthropic, Character.AI, etc.) experienced significantly degraded performance. Users reported slow, inconsistent responses, increased error rates, and intermittent unavailability—behaviors commonly seen during external network or storage disruptions.

• Real-time monitoring flagged elevated latency and 5xx errors from all LLM services.

• Incident triaging confirmed the root cause was an external internet/cloud infrastructure failure—not LLM model degradation.

• Fallback strategies (rate limiting, cached responses) were enacted where feasible until network stability returned later that afternoon.