Satya Nadella believes that if you run a company, you should build your own AI.
In an interview published Friday, Microsoft’s CEO told Applied Computing co-founder Yash Patil that all companies need to create AI models tailored to their business.
“I simply think there should be as many models as there are companies in the world,” Nadella said. “Because at the end of the day, what is a company? A company is a learning system.”
“I don’t want to be tied to one model,” Nadella added. “I want to be able to use my own context, my own data, and actually my own traces to get more open-weighted and cost-effective models or fine-tuned models.”
This comment represents one of Nadella’s clearest visions for enterprise AI yet. Many companies rely on foundational models from a smaller group of AI companies such as OpenAI, Anthropic, Google, and Meta.
Rather than relying solely on OpenAI, Microsoft is increasingly adopting a multi-model strategy through Azure AI Foundry, which also hosts models such as DeepSeek and Cohere. Amazon is pursuing a similar strategy with Bedrock, while Google Cloud offers a growing catalog of third-party and proprietary models alongside Gemini.
Many companies are also experimenting with open-weight AI models, such as Meta’s Llama and Mistral’s models, where parameters are exposed so companies can fine-tune and deploy the AI themselves.
Nadella said AI concentration poses long-term economic risks.
“You can’t say, ‘Hey, look, I have two frontier models, or I have three frontier models,’ or some finite set that has learned everything that is differentiated in today’s economy, because then the economy would collapse,” Nadella said.
“You can always buy tools, you can always outsource tasks and jobs, but you can’t outsource learning,” Nadella added. “If learning is outsourced, why does it exist?”