Can China’s DeepSeek AI Compete with US Tech Giants?

DeepSeek’s efficiency largely stems from reinforcement learning, allowing the model to improve itself with minimal human intervention. “This makes development cheaper, but it’s unclear whether the estimated $6 million cost also includes data acquisition,” Wiggers explains.

Despite its advancements, he sees limitations: “U.S. models support more languages and are better tailored for diverse markets. DeepSeek focuses primarily on Chinese and English, and research suggests it may produce more harmful outputs.”

Europe lags behind both China and the U.S., with just €1.5 billion in AI investments compared to the U.S.'s hundreds of billions. However, Wiggers sees a silver lining in DeepSeek’s emergence: “It demonstrates that significant progress can be made with fewer resources, which presents opportunities for European AI development.”

Watch the full interview here.