US Tech Giants Accuse Chinese AI Firms of Stealing Billions in Research
The AI cold war just got a whole lot hotter. US tech companies are now officially accusing Chinese AI firms of stealing billions in research, according to reports from Euronews on February 26, 2026. This is literally the kind of thing that gets called 'an AI cold war' in the headlines, so yeah, things are that tense.
OpenAI dropped a bombshell in February, telling US lawmakers that they caught DeepSeek - that's the Chinese AI company everyone's been talking about - trying to secretly copy their most powerful AI models. We're not talking about inspiration here, we're talking about alleged industrial-scale theft of cutting-edge research worth billions.
And it gets worse. Google and Anthropic both warned about something called 'distillation attacks' - basically when you take a powerful AI model and create a stripped-down version that doesn't have the same safety guards. The scary part? These distilled models apparently won't have the safeguards to prevent state and non-state actors from using AI in bioweapons or conducting cyberattacks. That's exactly the kind of stuff that keeps safety researchers up at night.
The Chinese companies? They're not exactly denying everything, but they're certainly not admitting to anything either. This whole situation is creating massive tensions between the two biggest AI powers in the world.
Here's why this matters: if Chinese companies are indeed stealing AI research, they're basically getting years of R&D investment for free. Meanwhile, US companies are spending billions on safety research that could literally save lives - and those safety features might be getting stripped out by distillation attacks.
The implications are huge:
- Billions in alleged IP theft from Chinese AI firms
- OpenAI caught DeepSeek trying to copy proprietary models
- Distilled models lack safety guards against bioweapons and cyberattacks
- Growing tension between US and China in the AI race
- Safety research potentially being undermined