DeepSeek Doesn’t Endanger Nvidia And U.S. Chip Export Controls: Fortune

AI has fueled Nvidia’s meteoric rise to a $3 trillion market valuation.

DeepSeek has disclosed that it possesses 10,000 of Nvidia’s older-generation A100 GPUs, which were acquired before U.S. export restrictions took effect. I Photo: EdTech Stanford University School of Medicine Flickr

But recently, AI also sparked panic among Nvidia investors, sending its shares plummeting by nearly 17% and wiping out almost $600 billion in value.

The selloff was triggered by Chinese AI startup DeepSeek, whose latest V3 and R1 AI models appear to rival those of any U.S. company—while being developed at a fraction of the cost, Jeremy Kahn reported for Fortune.

Because Nvidia’s powerful graphics processing units (GPUs) represent one of the biggest costs in developing cutting-edge AI models, investors are now reevaluating their assumptions about the AI industry.

However, predictions of Nvidia’s downfall may be premature.

Likewise, claims that DeepSeek’s success should prompt the U.S. to abandon its AI-related chip export controls may be overstated.

DeepSeek has disclosed that it possesses 10,000 of Nvidia’s older-generation A100 GPUs, which were acquired before U.S. export restrictions took effect. The company has also stated that it trained V3 using Nvidia’s H800 chips, a model specifically designed to comply with U.S. trade regulations.

Still, DeepSeek’s processing power pales in comparison to that of U.S. firms. For instance, Elon Musk’s AI venture, xAI, recently built a computing cluster in Tennessee called Colossus, equipped with 100,000 of Nvidia’s more advanced H100 GPUs.

Additionally, DeepSeek’s R1 model—designed to excel in math, logic, and coding—is small enough to run on a laptop, relying on a conventional central processing unit (CPU) rather than requiring large-scale GPU infrastructure.

Some experts believe that, counterintuitively, DeepSeek’s success could increase global demand for high-end AI chips, including those made by Nvidia and its competitors.

This phenomenon aligns with the Jevons Paradox, named after 19th-century economist William Stanley Jevons, who observed that when technological advances improve efficiency, overall consumption of a resource tends to rise.