Anthropic's infrastructure has explicit capacity management. When the number of concurrent users exceeds what Claude's backend can handle, new requests are rejected with this message rather than being queued indefinitely. This typically occurs during major product launches, news events, or peak daytime hours in the US. It is almost always a transient, provider-side issue that resolves within 15–30 minutes.