China has significantly reduced its AI development gap with the U.S. to just three months in some areas, thanks to firms like DeepSeek optimizing chip usage and algorithm efficiency, according to 01.AI CEO Lee Kai-fu.
DeepSeek made waves in January by launching an AI reasoning model trained with less advanced chips, challenging the notion that U.S. sanctions were hindering China’s AI sector. Lee, a former Google China head, noted that China had even surpassed the U.S. in infrastructure software engineering.
Washington’s semiconductor restrictions, he said, initially posed challenges but ultimately spurred innovation. Chinese firms have developed alternative reinforcement learning techniques, with DeepSeek’s model demonstrating advanced reasoning capabilities comparable to OpenAI’s.
China’s tech sector entered the generative AI race following OpenAI’s ChatGPT launch in 2022. While initially lagging, startups like DeepSeek, ZhipuAI, and Moonshot, along with giants Baidu, Alibaba, and ByteDance, have rapidly advanced.
Lee, who founded 01.AI in March 2023, highlighted the difficulty of competing with well-funded tech giants. Instead, his company focuses on practical AI applications, helping enterprises deploy AI solutions. Its new platform, Wanzhi, launched this month, has already generated revenue, with 2025 growth projections exceeding last year’s $15 million earnings.
Source: Reuters
Bd-pratidin English/FNC
