Microsoft Aims for AI Independence with Proprietary MAI Models

Microsoft CEO Satya Nadella has signaled a strategic shift in the company's AI approach. In an interview, he discussed why Microsoft hasn't yet developed a frontier model despite having a massive $175 billion AI budget.

Nadella emphasized that the tech giant is building its proprietary MAI models from scratch, unlike relying on external partners like OpenAI and Anthropic. He drew parallels to his early career working on Windows interoperability with Unix, where opening systems deepened Microsoft's enterprise reach rather than shrinking it.

Nadella noted that cloud hyperscalers cannot simply serve as passive computing suppliers to third-party model developers, calling this dynamic unsustainable. He revealed that Microsoft has already started rerouting workloads to its own MAI models, with promising early results in GitHub Copilot, Excel, and Outlook.

The company's goal is to achieve 'full independence' from external partners, with Nadella advising others to prioritize internal development over reliance on external models. He believes that software now has a real marginal cost for the first time, making the choice of model a business decision rather than a technical default.