Microsoft CEO Nadella warns: Companies using proprietary AI models give away knowledge
In a blog post, Microsoft CEO Satya Nadella warns that enterprises using proprietary AI models are paying twice – both in money and by handing over valuable proprietary knowledge that the models learn from their usage.
Microsoft CEO Satya Nadella published a blog post on Monday joining a growing chorus of voices warning about the risks of using proprietary AI models. He argues that companies using models from labs like OpenAI and Anthropic are not only paying for token usage but also unknowingly giving away valuable data.
Nadella points out that models learn from “exhaust” – the prompts users write, the tools agents use, and especially the corrections people make when the model is wrong. Every correction is distilled into institutional know-how that a competitor could never buy. He finds it hypocritical that model makers freely train on public data but then restrict others from distilling their models.
The solution Nadella proposes is typical of a large cloud provider CEO. He urges companies to retain ownership of their data, including prompts and feedback, and build their own “proprietary learning environments” in the cloud (which conveniently could be Microsoft Azure). He also recommends building “orchestration layers” that allow easy switching between different AI providers, much like the growing trend of AI gateways.
This trend is already happening. Idit Levine, CEO of Solo.io, says her customers, after experimenting with proprietary models, ask if they can use an open-source model on-premises, as it does almost 90% of the job at a fraction of the cost. Vercel and OpenRouter also report a surge in traffic to open-source models; last month, open models accounted for 29% of all traffic through Vercel’s gateway.
Nadella concludes: “In consuming intelligence, you are creating intelligence. And what you create should belong to you.”
