Microsoft CEO Satya Nadella is reinforcing a warning he first made earlier this month: companies that hand too much control to proprietary AI providers risk undermining their long-term competitiveness. He argued that enterprises should be far more deliberate about the data they share with AI model providers, from prompts and metadata to the business context that underpins internal workflows.

Why Data Control Matters

Nadella said companies should retain ownership of the information generated each time they use an AI model, allowing them to build their own capabilities over time.

Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram

“Every time you use the model, all of the metadata around it is retained by you, so that you could use all of that to train perhaps your own weights or your own open model,”

he said.

In AI, weights are the trained parameters that determine how a model behaves. Nadella argued that businesses should preserve enough operational data and metadata to develop or fine-tune their own models rather than becoming dependent on external providers.

“Any firm that doesn’t have this control, I will claim will not remain a firm because you’ve essentially outsourced your thinking,”

he added.

Build AI Infrastructure, Not AI Dependency

Nadella argued that enterprises should avoid locking themselves into a single AI provider and instead build infrastructure that allows them to switch between models as technology evolves.

That includes AI gateways that separate prompts, business context and memory from the underlying model. It also means keeping coding tools, interfaces and enterprise workflows independent of any single provider.

Nadella cited Anthropic’s Claude Code and OpenAI’s ChatGPT Codex as examples of tightly integrated coding tools that enterprises should avoid relying on exclusively.

“By keeping the harness separate from the model and the context and memory separate from the model, you absolutely can use multiple models for what they’re great at. At the same time, any one model can go away, and you can still continue to be in control of your own destiny,”

Nadella said.

A Strategy That Also Benefits Microsoft

Nadella’s comments align with Microsoft’s broader AI strategy. The company is an investor in both Anthropic and OpenAI while providing Azure cloud infrastructure that enables enterprises to deploy and manage multiple AI models.

The remarks also reflect a broader shift across the market. As open-weight models become more capable and cost-effective, enterprises are increasingly moving beyond a single-provider approach and adopting AI stacks that support multiple models.

That transition is driving demand for orchestration platforms, governance tools and model-agnostic coding agents capable of managing increasingly complex AI environments.

The Startup Risk Is Becoming An Enterprise Risk

Nadella’s warning extends beyond vendor lock-in. He argued that the more access companies give AI providers to proprietary workflows, internal systems and business knowledge, the greater the risk those providers could eventually compete with them.

The concern echoes a long-running debate in Silicon Valley, where startups have often worried that platform companies could study successful products before launching competing services.

In May, OpenAI CEO Sam Altman offered to invest in every startup in Y Combinator’s latest cohort through AI credits. Investor Jason Calacanis cautioned founders against becoming too dependent on a single platform.

“If you take these tokens, there’s a non-zero chance that OpenAI will study exactly what your startup is doing, copy your idea and put your app into their free offering,”

he posted.

Nadella is now extending a similar argument to large enterprises, suggesting that the platform risks long associated with startups are becoming increasingly relevant as companies deploy AI across core business operations.

Consumers Face A Different Trade-Off

Nadella distinguished enterprise users from consumers, arguing that concerns over proprietary business data do not apply in the same way to individuals using consumer AI services.

When asked how consumers should protect themselves, he described data sharing as part of the long-standing value exchange that underpins much of the consumer internet.

“To some degree there’s got to be some value exchange in the consumer space where you’re getting something for free, maybe for your data. That’s sort of how the advertising business model has worked,”

Nadella said.

For enterprises, however, Nadella’s message was clear: AI should remain infrastructure companies control, not intelligence they outsource.