DeepSeek's AI 'breakthrough' isn't what it seems

Chinese company's so-called efficiency gains don't eliminate need for massive AI compute

Avatar
20250213 deepseek ai

DeepSeek's success is built on top of a mountain of American-origin AI compute. © Sipa USA

Douglas Fuller is an associate professor of international economics, government and business at Copenhagen Business School and the author of "Paper Tigers, Hidden Dragons: Firms and the Political Economy of China's Technological Development."

Since its release last month, DeepSeek's open-source generative artificial intelligence model, R1, has been heralded as a breakthrough innovation that demonstrates China has taken the lead in the artificial intelligence race. Cue the hand-wringing in Washington D.C. and triumphalism in Beijing declaring that hardware export controls on chips enabling AI compute have been successfully circumvented through clever coding.

Sponsored Content

About Sponsored ContentThis content was commissioned by Nikkei's Global Business Bureau.