Microsoft CEO Satya Nadella wants every firm to develop its own AI model

Microsoft chief executive Satya Nadella has laid out an ambitious vision for the future of enterprise artificial intelligence, arguing that every business should eventually have its own AI model rather than depending entirely on systems developed by a small number of technology companies.

Speaking during an interview with Yash Patil, co-founder of Applied Compute, Nadella said organisations should treat AI as an extension of their own knowledge rather than as a one-size-fits-all tool. His remarks come as businesses across industries increasingly adopt generative AI to automate tasks, analyse data and improve productivity.

"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."

Businesses should own their AI

Nadella argued that companies gain a competitive edge through the knowledge they accumulate over time, making it important for them to build AI systems around their own data, context and operational experience.

More from Tech

Rather than becoming dependent on a single frontier model, he said businesses should have the flexibility to combine proprietary information with open-weight or fine-tuned AI models that are both cost-effective and adaptable.

"I don't want to be locked into any one model," Nadella said. "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 reflect Microsoft's evolving AI strategy. While the company remains a major partner of OpenAI, it has broadened the range of models available through Azure AI Foundry, which now supports offerings from developers including DeepSeek and Cohere alongside Microsoft's own AI services.

The approach mirrors a broader industry trend. Amazon's Bedrock platform provides access to multiple third-party models, while Google Cloud has expanded its catalogue to include both Gemini and external AI systems. At the same time, enterprises are increasingly experimenting with open-weight models such as Meta's Llama and those developed by Mistral, allowing organisations to customise AI for their specific needs.

Warning against AI concentration

Beyond advocating for custom AI, Nadella also warned against a future in which only a handful of companies control the world's most advanced AI models.

He suggested that excessive reliance on a small number of frontier systems could weaken competition and reduce the unique knowledge that individual businesses contribute to the wider economy.

"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.

He also stressed that while companies can outsource tools, services or even certain job functions, they cannot afford to outsource the process of learning itself.

"You can always buy a tool, you can even outsource a task or even a job, but you can't outsource your learning," Nadella said. "If you outsource your learning, then why exist?"

His comments underscore Microsoft's belief that the next phase of enterprise AI will be defined not only by increasingly powerful foundation models but also by how effectively organisations build AI around their own expertise, data and decision-making processes.