If you own a company, Satya Nadella thinks you should build your own AI. In an interview that went live Friday, Microsoft's CEO told Yash Patil, cofounder of Applied Compute, that every company should create AI models tailored to its own business.
"My simple thing is there should be as many models in the world as firms in the world," Nadella said. "Because after all, what is a firm? A firm is a learning system."
"I don't want to be locked into any one model," Nadella added. "I want to be able to use my own context, my own data — in fact, my own traces to maybe even take a much more open-weight, cost-efficient model or a fine-tuned model."
The comments mark one of Nadella's clearest visions yet for enterprise AI. Many companies rely on foundation models from a relatively small group of AI companies, including OpenAI, Anthropic, Google, and Meta.
Microsoft has increasingly embraced a multi-model strategy through Azure AI Foundry, which also hosts models such as DeepSeek and Cohere, rather than relying solely on OpenAI. Amazon has pursued a similar strategy with Bedrock, while Google Cloud offers a growing catalog of third-party and proprietary models alongside Gemini.
Many enterprises are also experimenting with open-weight AI models, which have publicly available parameters so that companies could fine-tune and deploy the AI themselves, such as Meta's Llama and Mistral's models.
Nadella said that AI concentration poses long-term economic risks.
"It can't be, 'Hey, look, I have two frontier models or three frontier models' or whatever, some finite set that have learned everything that is differentiated today in the economy because then it collapses,'" Nadella said.
"You can always buy a tool, you can even outsource a task or even a job, but you can't outsource your learning," Nadella added. "If you outsource your learning, then why exist?"