Microsoft Copilot is recovering after a major Azure-related disruption, but the service is not fully back to normal for everyone yet. Some users may still run into slow responses, unstable connections, or missing resources when trying to use the AI assistant.
The outage was not limited to Copilot. Microsoft said it was investigating an incident in Azure that affected multiple services at the same time, which helps explain why the chatbot was caught in the wider cloud problem.
User reports had climbed sharply while the disruption was active. Downdetector recorded more than 600 reports at one point, showing that the issue was affecting a broad group of users rather than a small number of accounts.
Microsoft’s status updates point to a technical trigger behind the incident. The company said the disruption was caused by widespread power outages after a severe thunderstorm, which led to higher latency, unstable connectivity, and timeouts when customers tried to reach resources.
That cloud-level problem also had a wider impact across Microsoft’s infrastructure. Services including Azure Functions, Azure Database for MySQL flexible servers, Azure Database PostgreSQL Flexible Server, Azure Databricks, Redis, Azure SQL, Azure Managed Grafana, Virtual Machine, Virtual Machine Scale Set, Azure Kubernetes Service, Storage, and Application Insights were also affected.
For Copilot users, the effect was straightforward but frustrating. The assistant could fail to open, respond slowly, or not load at all while the broader Azure disruption was underway.
Signs of recovery are now visible, and Downdetector shows Copilot no longer appears to be experiencing the same large-scale problem as before. Even so, Microsoft says some users may still face high latency, intermittent connectivity, or resources that are not yet available.
That means the experience can still vary from one user to another. Some people may already be able to use Copilot normally again, while others may continue to encounter occasional errors as recovery moves forward gradually.