Anthropic CEO Dario Amodei: DeepSeek’s Threat to AI Companies ‘Greatly Overstated’

Anthropic CEO Dario Amodei wrote in an essay posted Wednesday (Jan. 29) that claims about Chinese artificial intelligence (AI) company DeepSeek’s potential threat to U.S. AI leadership is “greatly overstated” and that the drop in chipmaker Nvidia’s stock driven by the release of the DeepSeek’s models was “baffling.”

In an essay primarily focused on making the case for export control policies on chips, Amodei said the release of the DeepSeek models is not bad for other AI companies and is not bad for Nvidia.

Amodei wrote that DeepSeek V-3, a model released about a month ago, appeared to come close to the performance of state-of-the art models on some important tasks, while costing less to train.

However, he wrote, a closer look shows that the DeepSeek model is close to the performance of U.S. models that are between seven and 10 months older; is “an expected point on an ongoing cost reduction curve”; and was trained by a company whose total spend as a company is not so different from U.S. AI labs.

Of DeepSeek’s R1, the model that was released last week and drew public attention to the company and drove down Nvidia’s stock, Amodei wrote that it’s possible for several companies to produce a model of this type.

“We’re therefore at an interesting ‘crossover point,’ where it is temporarily the case that several companies can produce good reasoning models,” Amodei wrote. “This will rapidly cease to be true as everyone moves further up the scaling curve on these models.”

PYMNTS reported Tuesday (Jan. 28) that BofA said DeepSeek’s claim that it could train a foundation model for only $5.58 million is “misleading” because DeepSeek did not include costs related to research, experiments, architectures, algorithms and data.

Making the case for export controls on chips to China, Amodei wrote in his Wednesday essay that these controls are a powerful tool for keeping democratic nations ahead of China in AI development.

“In the end, AI companies in the U.S. and other democracies must have better models than those in China if we want to prevail,” Amodei wrote. “But we shouldn’t hand the Chinese Communist Party technological advantages when we don’t have to.”

The post Anthropic CEO Dario Amodei: DeepSeek’s Threat to AI Companies ‘Greatly Overstated’ appeared first on PYMNTS.com.