For Chinese companies, it’s not about which AI is the smartest

This report is from CNBC’s The China Connection newsletter this week, delivering insights and analysis on the powerhouse of the world’s second-largest economy. You can subscribe here.

Will the US or China win the artificial intelligence race? That’s a big question for investors considering where to invest their money.

However, I frequently find that many companies in China are asking a different question. “What AI tools can help you survive in a tough economy?”

All of this means that investors who pick AI winners based solely on smarts may miss the bigger story.

First, there’s the cost factor.

According to Artificial Analysis, which benchmarks AI models, OpenAI’s gpt-oss openweight model and DeepSeek are the cheapest AI models to operate, at 30 cents per million tokens. Google’s Gemini 3 Pro costs $4.50, while Anthropic’s top-of-the-line Claude Opus 4.5 costs $10.

This is a pay-as-you-go model that can quickly inflate prices, whether companies are using OpenAI’s “closed” AI models or remotely accessing DeepSeek or Alibaba’s open source models.

Instead, many companies in China prefer to control costs by choosing the third option of downloading open source models, James Tong, CEO of Suzhou-based Movitech, told me. His enterprise software company’s clients include Starbucks and Unilever.

Another key difference with the US is how Chinese companies are leveraging AI.

According to Tong, companies in the US tend to use AI more for coding, while in China they are more likely to use AI for specific situations. For example, since late 2025, Movitech has been using AI agents to help manufacturers improve their production processes and help state-owned enterprises achieve compliance.

An open-source model that runs locally is a better fit because it doesn’t require users to upload information to a third-party cloud server.

This is a bonus productivity tool when domestic demand is weak.

Tong said economic pressures forced him to lower product prices, but AI has increased efficiency within the company. Without these benefits, the business would have been in a much weaker position, he said.

According to industry rankings such as Artificial Analytics, the gold standard for generative AI models still belongs to the United States, but these tools are not officially available in China.

Beijing’s firewalls have blocked Google and Facebook and their AI models from intrusion, and Anthropic cited national security concerns in banning Chinese-owned companies from using Claude, which is currently widely considered the best AI for coding.

However, discussions about the leading US model have been widespread enough on Chinese social media to suggest that such technology is being used by local developers. And conversations with various companies confirm that they are experimenting with different AI models, including one in the US, to find one that fits their needs.

“When choosing AI solutions, Chinese companies are no longer prioritizing ‘who is the smartest’,” said Sun Xin, vice president of research at Gartner. Instead, they seek authenticity, control over hallucinations, and the ability to integrate.

Improving productivity is especially important.

Thanks to AI, 10 people can now do the same amount of work as 20 to 30 people, said Wenhao Zhang, CEO of Beijing-based consumer marketing consultancy Doodod Technology.

As a result, the company’s customer count is about the same as last year, at well over 100 companies.

Low prices and ease of access mean Chinese-made models are gaining more users overseas, especially among those looking to try out the hot open source AI agent OpenClaw.

Chinese models also rank highly on OpenRouter, a popular platform that allows businesses to access multiple AI models through one portal.

Anthropic and Google’s U.S. models were still the top two in usage last week, but DeepSeek’s model took third place, and fellow Chinese Moonshot AI’s Kim’s K2.5 took fourth place just a few days after its launch.

Moonshot said that since the release of its latest K2.5 model, new overseas users have increased fourfold, and revenue from outside China has exceeded revenue from the domestic market.

Reflecting investors wanting a piece of the pie, CNBC reported last month that Moonshot’s valuation soared by at least $500 million in just a few weeks after rivals Zhipu and Minimax went public in Hong Kong.

As the Lunar New Year approaches in China, Chinese companies and technology developers are hoping that DeepSeek will release another groundbreaking AI update. Just before the holiday season last year, the Chinese company released an open source model that costs far less to use than OpenAI’s ChatGPT.

DeepSeek’s release at the time shocked global investors who thought China could not innovate in AI without high-end Nvidia chips. Local companies previously locked out of major U.S. AI models have begun experimenting with DeepSeek and similar open-source models developed by Alibaba, ByteDance, and local startups.

“As long as China’s private sector continues to have revenue capacity and China’s innovation can catch up with the U.S., people will be interested in Chinese stocks,” said Ren Lijian, director of modern alpha at U.S.-based fund manager WisdomTree.

As China’s AI models try to catch up with their American rivals, innovation isn’t just about being intelligent, it’s also about reaching a business-ready price point.

Damian Ma, director of Carnegie China, said we now live in an “a-la-carte, profit-based world” rather than a values-based world, and he doubted there would be a new “golden age” for UK-China and US-China economic relations.

Zhiwei Zhang, president and chief economist at Pinpoint Asset Management, said the market expects policy changes in the real estate market from the fourth quarter of 2025 onwards, but questions remain as to whether the reported relaxation of borrowing rules for developers will be enough to change momentum in the sector.

Chinese stocks were little changed on Wednesday amid mixed regional trading, but software stocks took a hit as investors weighed the threat of AI automating workflows, squeezing prices and lowering barriers to market entry for competitors.

hong kong Hang Seng Index As of 12:45 a.m. ET, CSI 300 was largely unchanged. Year-to-date, both indexes are up about 4.6% and 0.7%, respectively.

However, the Hang Seng High-Tech Index fell by about 1.8%. The offshore yuan was little changed at 6.933 yuan against the dollar.

Chinese software companies fell along with their peers in other regions. Shares of China’s KingD International Software fell more than 15%, while cloud giant Tencent fell 3.27%. alibaba fell more than 1%, and Baidu fell more than 2%.

February 7: Uruguay President Yamandu Orsi ends nearly week-long state visit