The explosion of "cheap" alternative to ChatGPT is the result of China’s bet on internal competition among large companies and startups to develop the most advanced AI systems. A stinging defeat in a Go tournament is having a greater impact than US sanctions on the advanced microchip exports to China.

Milan (AsiaNews/Agency) – The performance of the new Chinese artificial intelligence model DeepSeek V3 and its competition with ChatGPT and the giants of the sector have sparked a lot of talk. The same goes within China, as Chinese companies compete with each in this strategic sector of the world economy.

Alibaba, China’s e-commerce giant, has reportedly a new version of its Qwen 2.5 artificial intelligence model, deemed better than DeepSeek-V3, which was hailed for its lower costs and energy consumption compared to Western models (despite or perhaps because of the ban on the sale of advanced chips to Chinese companies).

“Qwen 2.5-Max outperforms ... almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B," Alibaba's cloud unit said in a post on its official WeChat account, referring to OpenAI and Meta's most advanced open-source AI models.

Qwen 2.5-Max’s launch took place in parallel with cloud computing technical support for the live broadcast of the Spring Festival Gala, the People's Republic of China's most important annual variety show, which mixes music, dance, opera, martial arts and comedy.

On 12 January, two days after the release of DeepSeek-R1, ByteDance, the company that owns TikTok, released an update to its flagship AI model, claiming to have passed OpenAI's o1 version (backed by Microsoft) in AIME, a benchmark test that measures the ability of AI models to understand and respond to complex instructions.

The race among Chinese models with the entry on the market of subjects unknown until yesterday such as DeepSeek is not a sudden event, but the result of a precise strategy implemented by Beijing.

Chase Young, a young scholar, explained this a few weeks ago in an interesting article published on the website of the Cornell University.

“Following the success of ChatGPT and restrictive U.S. sanctions on chips, most expected China to concentrate its resources to compete in the AI space,” he writes. “Instead, Beijing is pursuing a strategy being referred to within China as the “war of [a] hundred models.”

“The strategy appears to be similar to China’s strategy in EVs, where it offered a wide array of subsidies. The EV strategy resulted in impressive industry leaders such as BYD and Li Auto, but also a glut of over 200 EV manufacturers, many of which are unprofitable. Similarly, many of China’s AI startups are currently facing financial difficulties.”

In this context, Young says, last summer, at the World Conference on Artificial Intelligence in Shanghai, Baidu's CEO, Robin Li Yanhong, expressly raised the question by asking whether China does not have too many artificial intelligence startups today and whether it would not be more fruitful to concentrate resources.

DeepSeek's exploit seems to suggest that “one hundred models” will continue, also because Chinese artificial intelligence start-ups can count on important advantages.

Young cites data from the Organisation for Economic Co-operation and Development (OECD) showing that China currently contributes more than 20 per cent of all AI research in the world, more than the European Union and India combined.

China also dominates the global race for generative AI patents, with “six times more than second-place United States” in the last 10 years.

In the article, Young also identifies a turning point in Chinese policies on the matter, linked to a curious competition China lost.

“The ‘Future of Go’ summit in May 2017 is commonly seen as the genesis for China’s ‘New Generation Plan’,” followed two months later by the claim that “China will become ‘the world’s primary AI innovation center’.”

What happened? During the summit, Google's AlphaGo artificial intelligence programme defeated five top-tier Chinese Go players.

“Google did not plan on spurring massive Chinese investment in AI,” Young explains. However, “for China, having its top players in its own national pastime defeated by an American company was seen domestically as a ‘Sputnik[*] Moment’.”

“Beyond investing at the university level, in November 2017 China began tasking Baidu, Alibaba, Tencent, and iFlyTek with building ‘open innovation platforms’ for different sub-areas of AIs, establishing them as national champions for the AI space. All four continue to invest in AI models today and the program has grown to at least 15 companies.”

RED LANTERNS IS THE ASIANEWS NEWSLETTER DEDICATED TO CHINA. WOULD YOU LIKE TO RECEIVE IT EVERY THURSDAY? TO SUBSCRIBE, CLICK HERE.

[*] Sputnik was the first artificial Earth satellite. It was launched into an elliptical low Earth orbit by the Soviet Union on 4 October 1957 as part of the Soviet space program. I

• china
• artificial intelligence
• ai
• competition
• united states