Technology

DeepSeek reduces the China-US AI gap to just three months, claims 01.AI founder Lee Kai-fu.

Published On Tue, 25 Mar 2025
Aditya Patil
2 Views
news-image
Share
thumbnail

BEIJING/HONG KONG, March 25 (Reuters) – China's artificial intelligence (AI) sector has significantly reduced its gap with the United States, narrowing it to just three months in certain areas, according to Lee Kai-fu, CEO of Chinese startup 01.AI. This progress is attributed to companies like DeepSeek, which have optimized chip usage and algorithm application.

Lee, a well-known AI expert and former Google China head, told Reuters that DeepSeek’s advancements have positioned China ahead in fields such as infrastructure software engineering. The company made headlines in January when it introduced an AI reasoning model developed using less advanced chips at a lower cost compared to its Western counterparts, challenging the belief that U.S. sanctions were stalling China’s AI progress.

"In the past, China lagged behind the U.S. by six to nine months across all aspects. Now, the gap in core technologies has shrunk to around three months, and in some areas, China has taken the lead," Lee said in an interview in Hong Kong. He described Washington’s semiconductor restrictions as a "double-edged sword," presenting short-term obstacles while pushing Chinese firms to innovate. He highlighted that DeepSeek’s breakthroughs in reinforcement learning have either helped China catch up quickly or even surpass the U.S. in certain innovations. The company's AI models demonstrate reasoning processes before delivering answers, a feature pioneered by OpenAI but not yet made available to users.

Following OpenAI’s launch of ChatGPT in late 2022, China’s tech industry quickly joined the race to develop generative AI. However, before DeepSeek’s emergence, many Chinese tech leaders acknowledged that they lagged behind their Western counterparts. Lee, who also leads a venture capital firm, founded 01.AI in March 2023. His company joins other Chinese AI startups like ZhipuAI and Moonshot, alongside tech giants such as Baidu, Alibaba, and ByteDance, in developing foundational AI models.

He noted that investing in proprietary models has become increasingly challenging for startups due to competition from large tech firms and rapidly evolving open-source alternatives. Instead, 01.AI is focusing on practical AI applications, providing software solutions to help businesses effectively implement foundational models. Earlier this month, 01.AI introduced Wanzhi, a software platform designed to assist enterprises in deploying AI technology. The company has already begun generating revenue and expects significant growth in 2025, forecasting several times the $15 million earned last year, Lee added.


Disclaimer: This image is taken from Reuters.