Over 20 tech giants including Nvidia, Microsoft, Meta and Palantir released a joint letter Friday urging policymakers to avoid "premature restrictions" on AI distillation techniques that could "stifle competition or drive innovation overseas."

The coordinated response follows mounting concerns about Chinese lab Moonshot AI's Kimi K3 model, which users found competitive with leading offerings from Anthropic and OpenAI despite being released as an open-weight model.

White House advisor Michael Kratsios accused Moonshot of distilling Anthropic's Fable model to develop K3. "We have information that Moonshot AI distilled Anthropic's Fable for the development of its K3 model," Kratsios posted on X Wednesday, claiming the company developed "a sophisticated internal platform to conduct large scale distillation against U.S. models."

Distillation involves using outputs from advanced AI models to train other models. Critics argue it allows competitors to create offerings by leveraging billions of dollars in research investments made by frontier labs.

"It's almost like someone went to the lectures, read the textbook, and did all the hard work of doing the homework," said Pukar Hamal, founder of AI security firm SecurityPal. "Then some other student is like, 'Hey, I didn't do that. Can I just copy your work?'"

Why the technique matters

Google AI lead Jeff Dean discussed distillation on a podcast in February, explaining how his team used the technique to improve performance without relying on large models. "Through distillation, which is a key technique for making the smaller models more capable, you have to have the frontier model in order to then distill it into your smaller model," Dean said.

The practice has become widespread across the industry. Nvidia used distillation for its Llama Nemotron model series, according to research papers. Many companies incorporate the technique when creating their own models, said Shashi Bellamkonda, research director at Info-Tech Research Group.

Box CEO Aaron Levie, a signatory of Friday's letter, argued U.S. companies need access to the best technology regardless of origin to stay competitive. "Generally the arc is going to be that the more innovation that there is, whether that's from the U.S. or China or otherwise, you should expect more AI progress," Levie said.

The debate highlights tensions between protecting American AI investments and maintaining open innovation practices that have driven industry progress.