On Sunday, Microsoft CEO Satya Nadella doubled down on the shocking warning he issued earlier this month to companies using AI, and this time he went a step further. Companies that rely entirely on proprietary AI labs for their AI needs ultimately won’t survive, he predicts.

That’s what he said on CNN. Fareed Zakaria GPS. When Zakaria asked Nadella to explain what constitutes a company sharing too much with an AI model vendor, Nadella said companies need to be careful about everything they hand over, from their data to their prompts.

Nadella called for a setup where “every time you use the model, all the metadata around it is retained by you, so you can use all of it to train maybe your own weights or your own open model.” (Weights are the trained parameters of a model, essentially its brain. Nadella’s point: Companies should retain their own usage data so they can eventually build their own model.)

“I will say that any company that doesn’t have this control will not remain a company because they have essentially outsourced their thinking,” he added.

Bottom line: Companies without their own models (or without a layer of AI infrastructure known as AI gateways to separate their cues from the model itself) will be in trouble, Nadella says.

Specifically, he wants companies to stop relying on AI labs’ built-in coding tools, known as harnesses (Anthropic’s Claude Code and OpenAI’s ChatGPT Codex are examples of this).

“By keeping the harness separate from the model and the context and memory separate from the model, it is absolutely possible to use multiple models for what they are excellent at. At the same time, any model can disappear, and you can still remain in control of your own destiny,” Nadella said.

Of course, Microsoft is an investor in the two largest artificial intelligence laboratories, Anthropic and OpenAI. Coding agents are a particularly popular way for companies to use AI models, and by all indications, they are making big money for model makers.

And yet, Nadella tells companies not to rely too much on them. Microsoft would naturally benefit from that warning, since its cloud business is now also dedicated to selling the kind of alternative infrastructure it recommends.

Despite the obvious selfish scare tactic, he’s not wrong. Companies are increasingly realizing that they need many model options, particularly cheaper options, and are turning to open models (models whose underlying code is publicly available) that they can fine-tune and run on their own hardware. That, in turn, means they will also need ways to manage multiple models, as well as coding agents that are not tied to a specific model provider.

But Nadella’s observation doesn’t just refer to runaway budgets. He anticipates that once a company has “outsourced its thinking” to a model, there is little to stop the AI ​​lab from eventually offering a competing service of its own. This risk grows as companies adopt AI agents and give them access inside the company.

It’s the kind of warning that has rattled the startup industry for years: What’s to stop model makers from taking down startups by copying and competing with them?

In May, for example, when OpenAI CEO Sam Altman offered to invest in all of the Y Combinator startups in his latest group by offering them AI credits, investor Jason Calacanis issued a similar message: “If you take these tokens, there is a non-zero chance that OpenAI will study exactly what your startup is doing, copy your idea, and put your app in their free offering. This is the classic platform playbook: be careful, founders!” public.

Now Nadella is making the same argument to companies.

One caveat: Nadella’s concern about overusing AI models applies only to companies, not individuals. When Zakaria specifically asked Nadella how everyday people could protect themselves, Nadella downplayed it and said that data sharing is simply the price consumers pay for using a service, especially a free one.

“To some extent there has to be some value exchange in the consumer space where you get something for free, maybe for your data. That’s how the advertising business model has worked,” Nadella said.