Microsoft has rolled out its own in-house AI models at aggressively low pricing, a direct move to cut reliance on third-party labs and squeeze API costs for anyone building on Azure. The play lands as India becomes one of the fastest-growing AI developer markets, and it resets the math for Indian startups that have been burning runway on per-token bills. Cheap frontier capability from a hyperscaler means the build-versus-buy decision just got harder, and the margin game just got sharper.
⚡ Fast Takeaways:
- Core Update: Microsoft launched its own in-house MAI-family models across Azure AI Foundry, reducing dependence on OpenAI for core workloads.
- Key Metrics / Specs: Pricing lands well below comparable frontier APIs, with Microsoft positioning the stack for high-volume, cost-sensitive inference rather than benchmark chasing.
- Access & Availability: Rolling out through Azure AI Foundry and Microsoft Foundry endpoints, available to developers and enterprises now, with India regions in scope.
- Why It Matters Locally: Indian SaaS and AI product startups can now renegotiate token economics without switching clouds.