Is China quietly outpacing OpenAI in the race to AGI?

The AI race just got personal.

China's DeepSeek claims they’ve developed their R1 language model to rival OpenAI but they say they did it with just hundreds of GPUs, compared to the tens of thousands OpenAI uses. 

If true, this means DeepSeek has figured out how to train competitive AI models with far fewer resources - a major leap in efficiency.

But this might not be entirely true. Rumors suggest they could be hiding the fact that they have access to far more GPUs.

U.S. export controls prevent China from buying Nvidia’s top-tier H100 chips, forcing companies like DeepSeek to use downgraded A800 GPUs. However speculation is growing that DeepSeek stockpiled these A800s in massive quantities or even found ways to bypass the restrictions altogether. If that’s the case, their "low-resource" narrative might be more of a PR move than reality.

Either way this is a wake-up call for the industry. A reminder that the AI race isn’t just about who has the most GPUs anymore...

AI is becoming the power of the next frontier, and everyone is fighting to be number one.