Satya Nadella quietly criticized AI labs like Anthropic for the way they train their models.

In an X post on Sunday, Microsoft’s CEO said that model makers complaining about distillation are hypocritical. Distillation is the process of training a less powerful model based on the results of a stronger one.

“While the great innovation that comes from model vendors having fair use rights to train models with public data is needed, I find it ironic that the status quo is to then turn around and impose restrictive terms on distillation, and reserve the right to learn from customer usage and interaction data,” Nadella wrote.

He added that if learning only flows in one direction, the owners of the learning infrastructure will make all the money, while the knowledge creators will be excluded.

Frontier AI model builders like Anthropic, OpenAI, and Google DeepMind rely on work created by others to train their own models. ChatGPT, Claude and Gemini acquire their “intelligence” from writings, images and other publicly available data. Numerous companies and individuals have sued leading AI labs for scraping non-consensual content.

Although the lab was not named, Nadella’s comments seemed especially directed at Anthropic. Earlier this year, Anthropic CEO Dario Amodei complained that Chinese model makers are stealing his company’s work and using Claude to train their own models.

Last month, Anthropic wrote a letter to South Carolina Senator Tim Scott and Massachusetts Senator Elizabeth Warren, saying that Alibaba had recently carried out “the largest known distillation attack” to date.

“Competitors can use it to acquire powerful capabilities from other laboratories in a fraction of the time, and at a fraction of the cost, it would take to develop them independently,” Anthropic said in a lengthy statement on the topic in February.

Alibaba did not publicly respond to Anthropic’s allegations at the time.

In Sunday’s blog post, Nadella warned that companies that rely on leading models are essentially handing over their proprietary data and then paying to use it.

He said companies should own their AI infrastructure and institutional knowledge rather than relying on a single modeling vendor. They should also conduct their own assessments and their own “learning cycle,” allowing their AI capabilities to continually improve over time.

“That’s why companies need a real trust limit so that their human capital and their symbolic capital combine,” he said. “And it is a hard border that nothing crosses, not even exhausted intelligence, without consent.”

Elon Musk also criticized Anthropic for how it collects data and trains its models.

“Anthropic is guilty of stealing training data on a massive scale and has had to pay multi-million dollar settlements for its theft. This is just a fact,” Musk wrote in a February X post, following Anthropic’s complaint against the Chinese models.

Anthropic did not immediately respond to a request for comment from Business Insider.