Microsoft CEO Nadella Warns Enterprises Using Proprietary AI: ‘You Pay for Intelligence Twice’

Microsoft CEO Satya Nadella published a striking blog post on Monday that has sent ripples through Silicon Valley, warning enterprises that using proprietary AI models carries a hidden danger far greater than the cost of tokens. The post addresses what Nadella calls a growing concern among AI enthusiasts: that giant AI labs selling proprietary models are acting like “Trojan horses” inside the enterprises that use them.

“You essentially pay for intelligence twice, once with money, and again with something even more valuable: the proprietary knowledge you must reveal to make that intelligence useful,” Nadella writes. “The better you want the model to perform, the more of that knowledge you have to feed it!”

The concern is that as enterprises feed sensitive business data into models from OpenAI, Anthropic, and others, these labs gain ever-increasing access to their customers’ most confidential information. The model makers could then use that knowledge for themselves, potentially becoming competitors to their own customers. Nadella points out that “models learn from ‘exhaust’ — the prompts people write, the tools agents use, and especially the corrections people make when the model is wrong.” He calls this “the kind of knowledge a competitor could never buy.”

Nadella also takes aim at the asymmetry in AI data practices. He argues it is hypocritical for model makers to freely scrape the public internet to train their models while imposing restrictive terms on “distillation” — the practice of using a model’s outputs to study its behavior and train competing models. In February, Anthropic accused Chinese open-source models of using Claude’s outputs to improve their own systems and urged the U.S. government to tighten export controls. Nadella’s point is that model makers cannot have it both ways.

As CEO of Microsoft — itself a majority stakeholder in OpenAI and also an investor in Anthropic — Nadella is in a unique position to issue such a warning. His proposed solution is predictably cloud-friendly: he urges companies to “retain ownership” of their data, including prompts and correction feedback, and to build “proprietary learning environments” on the cloud — likely Microsoft Azure. He also advocates for “orchestration layers” that allow seamless switching between AI models from different providers, avoiding vendor lock-in.

The warning arrives as evidence mounts that enterprises are already shifting toward open-source models running on their own infrastructure. Idit Levine, founder and CEO of Solo.io, whose customers include T-Mobile, ADP, and SAP, told TechCrunch that after experimenting with proprietary models, her clients increasingly ask: “Can I take an open source model and run it on-prem? It will do almost 90% of what the big one’s doing. It will cost way less, and they can control it.” Vercel reports that open models accounted for 29% of all traffic routed through its gateway in June.

“In consuming intelligence, you are creating intelligence. And what you create should belong to you,” Nadella concludes. With Microsoft’s CEO now openly urging caution around the very AI labs his company has backed, this trend toward open-source, on-premise AI deployment is likely to accelerate significantly. The question for enterprise decision-makers is whether the convenience of proprietary models is worth the long-term cost of data sovereignty.