Microsoft and OpenAI restructured their partnership: OpenAI can now use any cloud provider, and the AGI clause defining their deal is gone.
Microsoft announced major changes to its OpenAI partnership on Monday. OpenAI retains Microsoft as its 'primary cloud partner' with Azure first-rights, but can now sell products via any cloud provider including AWS or Google Cloud. Microsoft retains a revenue cut from non-Azure deals. The companies also eliminated the AGI clause that previously governed what would happen if either party achieved artificial general intelligence — a clause that had significant contractual weight over the relationship's future.
The Azure lock-in on OpenAI APIs is officially loosening. If you've been routing OpenAI calls through Azure specifically because it was the only serious tier-1 path, that constraint no longer holds structurally. OpenAI can now negotiate native integrations with AWS Bedrock or Google Vertex — which means latency profiles, SDK availability, and billing integrations could shift within months as OpenAI courts enterprise on new infrastructure.
If your stack runs on AWS or GCP, benchmark OpenAI's current API latency directly via openai.com against your Azure-hosted calls this week — document the delta so you're ready to migrate routing the moment a native integration ships.
Suggested exercise · not independently tested. Access requirements and costs may vary.
Run: curl https://api.openai.com/v1/chat/completions -H 'Authorization: Bearer $OPENAI_API_KEY' -H 'Content-Type: application/json' -d '{"model":"gpt-4o","messages":[{"role":"user","content":"ping"}]}' -w '\nTime: %{time_total}s\n'
Run the same prompt routed through your Azure OpenAI endpoint and record both response times
Compare the two latency numbers and note which cloud path is faster for your region
Tags
Related