Microsoft quietly searches for OpenAI alternatives

Microsoft, which invested $13bn in OpenAI, is quietly reviewing AI startup acquisitions and partnerships to reduce its dependence on OpenAI just three weeks after the contract was revised. The April 27 revision ended the exclusive licensing of OpenAI models, allowing OpenAI to sell its models to other clouds, including AWS. In return, Microsoft retained an IP license through 2032, a 27% stake (about $135bn as of the last disclosure), and a clause giving Azure priority deployment of new OpenAI products.

The candidates Microsoft has examined fall into two categories.

• Cursor: Microsoft reviewed an acquisition plan for the code-generation startup, but dropped it due to heightened regulatory conflicts from holding it alongside GitHub Copilot. Afterward, SpaceX-xAI effectively locked up Cursor by attaching a $60bn option and a $10bn termination fee.
• Inception: A Stanford spinoff startup developing a diffusion-based language model instead of an autoregressive one. It processes tokens in parallel and claims speeds of over 1,000 tokens per second; Microsoft's M12 already participated in a $50m round last November.

This strategy is linked to the MAI Superintelligence team led by Mustafa Suleyman. The team launched in November 2025 and released MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2 in April. The target timeline for a general-purpose frontier LLM is 2027, per Suleyman's March memo. Ultimately, what Microsoft is looking for is a fallback option around code generation, model architecture, and developer touchpoints, rather than the AGI race itself.