Microsoft Chief Executive Satya Nadella has outlined a vision for enterprise artificial intelligence in which every company develops AI models tailored to its own business, arguing that organizations risk losing their competitive edge if they rely solely on models built by a handful of technology giants.

Speaking in an interview with Applied Compute co-founder Yash Patil, released on Friday, Nadella said AI should become a core capability of every enterprise rather than a service entirely outsourced to external providers.

“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.”

The comments provide one of the clearest articulations yet of Microsoft’s long-term enterprise AI strategy, one that emphasizes customization rather than dependence on a single foundation model.

Today, most companies deploying generative AI rely on frontier models developed by a relatively small number of firms, including OpenAI, Anthropic, Google, and Meta. While these large language models have accelerated enterprise AI adoption, Nadella argued that businesses should increasingly build on top of them using their own proprietary data, workflows, and institutional knowledge.

“I don’t want to be locked into any one model,” he 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.”

Microsoft has been pushing a growing shift toward a multi-model ecosystem. Although the software giant remains one of OpenAI’s biggest strategic partners, it has expanded Azure AI Foundry into a platform that supports a wide range of AI models, including offerings from DeepSeek, Cohere, and other developers, alongside OpenAI’s models.

The approach mirrors a broader trend among hyperscale cloud providers. Amazon’s Bedrock platform offers customers access to multiple foundation models, while Google Cloud has expanded its catalog beyond its proprietary Gemini family to include third-party alternatives.

The strategy is also part of the growing enterprise interest in open-weight AI models, whose publicly available parameters allow organizations to fine-tune systems using proprietary corporate data while maintaining greater control over costs, performance, and deployment. Models such as Meta’s Llama and those developed by Mistral have become increasingly popular for this purpose.

It is believed that encouraging customers to develop customized AI models also strengthens Microsoft Azure’s position as the infrastructure layer powering enterprise AI, regardless of which underlying model customers choose.

Beyond technology choices, Nadella framed the issue as one of economic competitiveness. He warned that allowing a handful of frontier AI developers to accumulate most of the world’s valuable knowledge would undermine innovation across industries.

“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.

His argument lends credence to a growing debate within the AI industry over whether value will ultimately accrue to companies building the largest frontier models or to enterprises that use those models to create proprietary, domain-specific intelligence. Rather than viewing AI as a commodity purchased off the shelf, Nadella suggested companies should treat it as an extension of their own institutional learning.

“You can always buy a tool, you can even outsource a task or even a job, but you can’t outsource your learning,” he said. “If you outsource your learning, then why exist?”

Currently, businesses worldwide are accelerating investment in generative AI while increasingly seeking ways to differentiate themselves beyond simply deploying the same commercially available models as competitors. This has made Nadella’s opinion profound, especially as Microsoft has been positioning Azure not merely as a cloud provider, but as the platform where enterprises can build, fine-tune, and operate AI systems that reflect their own data, expertise, and competitive advantages.

Nadella’s vision also signals that the next phase of enterprise AI may shift from a race to access the most powerful foundation models toward a race to build proprietary AI capabilities that embed each organization’s unique knowledge, processes, and decision-making into customized models.