How is China going to conquer the AI ​​market in the United States?

These are different models from the closed generative AI models that have become household names – such as OpenAI’s ChatGPT, or Google’s Gemini – whose ins and outs are jealously protected.

In contrast, the “open” models offered by many Chinese rivals, from Alibaba to DeepSeek, allow programmers to customize bits of software to suit their needs.

Globally, the use of open models developed by Chinese has gone from just 1.2% at the end of 2024 to around 30% in August of this year, according to a report published this month by the developer platform OpenRouter and the American venture capital firm Andreessen Horowitz.

Chinese open models “are cheap, in some cases free, and work well,” Wang Wen, dean of the Institute of Financial Studies at Renmin University of China, told AFP.

An American entrepreneur, who spoke on condition of anonymity, told AFP his business saves $400,000 annually by using Alibaba’s Qwen AI models instead of private models.

“If you need cutting-edge capabilities, you go back to OpenAI, Anthropic or Google, but most applications don’t require it,” the entrepreneur explained.

US chip giant Nvidia, AI firm Perplexity and Stanford University in California are also using Qwen models in some of their work.

The impact of DeepSeek

The January launch of DeepSeek’s high-performance, low-cost, open-source “R1” language model challenged the perception that the best AI technology had to come from American giants like OpenAI, Anthropic or Google.

LOOK: We will never be able to know if AI becomes conscious, according to a philosopher from the University of Cambridge

It was also a reckoning with the United States, locked in a battle for AI dominance with China.

The US government is aware of the potential of open source. In July, the Trump administration released an “AI Action Plan” that said the United States needed “leading open models based on American values.”

These models could become global standards, he noted. But so far, American companies are taking the opposite path.

Meta, which had led the country’s open source efforts with its Llama models, is now focusing on closed source AI.

However, this summer OpenAI—under pressure to revive the spirit of its nonprofit origins—launched two models known as “open weights,” that is, slightly less malleable than “open source.”

“Western ideas”

Among large Western companies, only France’s Mistral remains loyal to open source, but it ranks far behind DeepSeek and Qwen in usage rankings. Western open source offerings “are just not that interesting,” said the American entrepreneur who uses Alibaba’s Qwen.

LOOK: Severe anxiety and obsessive-compulsive disorder: ChatGPT, Grok, Gemini and other AI tools underwent psychological evaluation

The Chinese government has encouraged open source AI technology, despite questions about its profitability.

Mark Barton, chief technology officer at OMNIUX, said he was considering using Qwen, but some of his customers might be uncomfortable with the idea of ​​interacting with a Chinese-made AI, even for specific tasks.

Given the current US administration’s position on Chinese technology companies, risks remain, he told AFP.

“We wouldn’t want to go all in on a single model supplier, especially one that may not be aligned with Western ideas,” Barton said.

“If Alibaba were sanctioned or its use was, in practice, blacklisted, we do not want to be caught in that trap,” he summarized.

A recent Stanford study posited that “the very nature of open model publications allows for better scrutiny” of the technology.

Gao Fei, chief technology officer at Chinese AI wellness platform BOK Health, agrees.

“The transparency and shared nature of open source are in themselves the best ways to build trust,” he said.