Microsoft CEO Satya Nadella is warning companies about the risks of using AI models from major labs like OpenAI and Anthropic, suggesting that these models act like 'Trojan horses.' The concern is that while companies pay for AI token usage, they also inadvertently provide valuable business data to the AI providers.
Double Payment for Intelligence
In a blog post published Sunday, Nadella explained that AI users pay twice for intelligence: first with money, and then with proprietary knowledge they must share to make the AI effective. 'The better you want the model to perform, the more of that knowledge you have to feed it,' he notes. This information, he argues, could give AI providers a competitive edge over their own customers.
Distillation and Model Knowledge
Nadella advocates for a more balanced approach, proposing that enterprises should have the right to 'distill' AI models, a process of using a model's outputs to train new models. He criticizes the current state where model providers can train on public data but impose restrictions on others attempting the same with their models. 'It's ironic to impose restrictive terms on distillation,' Nadella writes.
Data Ownership and Switching Models
Nadella's solution includes companies retaining ownership of their data and building proprietary learning environments, potentially on cloud platforms like Microsoft's Azure. He also suggests developing 'orchestration layers' to easily switch between different AI models, avoiding dependency on a single provider.
Shift Towards Open Source Models
Idit Levine, CEO of Solo.io, notes a trend where companies are increasingly adopting open-source models on their premises. 'They understand that, and they can control it,' she states. Vercel and OpenRouter are also experiencing a rise in traffic to open-source models, with open models comprising 29% of Vercel's routed traffic last month.
Nadella's warning, coming from a company that invests in AI labs like OpenAI and Anthropic, underscores the growing movement towards open-source AI solutions. 'In consuming intelligence, you are creating intelligence. And what you create should belong to you,' Nadella concludes.
Source: https://techcrunch.com/2026/07/13/satya-nadella-has-issued-a-shocking-warning-to-companies-using-ai/
