Chinese open-weight AI models (models with freely downloadable parameters) now trail their American counterparts by just six to 12 months, according to Kai-Fu Lee, the AI entrepreneur and former president of Google China. Lee delivered that estimate in an interview with Fortune Daily published this week, and he argued that U.S. export controls on advanced chips have backfired. Instead of holding China back, the restrictions pushed Chinese developers to build competitive models with far fewer resources.

Lee used the interview to dismantle what he called three of the biggest misconceptions Americans hold about China's AI industry. His first target: the belief that Chinese companies care more about winning the AI race to artificial general intelligence (AGI, the hypothetical point at which AI matches human capability across tasks) than about building profitable businesses.

Myth one: Chinese AI firms are racing for AGI, not profit

"Chinese companies are born to make money. That's the primary purpose," Lee said. "It's not so much to win AGI or win the race."

He drew a direct comparison between questions surrounding the profitability of today's Chinese AI developers, such as DeepSeek, and the early skepticism that once met Pinduoduo, Meituan, and ByteDance. Those three companies faced similar doubts about their business models before growing into some of China's largest technology firms.

"You may reasonably doubt China's ability to do a breakthrough, which is valid. [The] U.S. is way ahead on that, but never doubt Chinese ability to find a way to make money," Lee said.

Myth two: Open-source adoption is a government project

The second misconception, according to Lee, is that China's embrace of open-weight and open-source technology forms part of an organized government effort to compete with the United States. He rejected that framing. Market competition, not state direction, drove the shift, he said.

Lee pointed specifically to DeepSeek, the Chinese startup that stunned American AI companies in early 2025 by releasing a highly capable model of the same name, which it claimed required significantly less computing power to train than its American competitors.

"Chinese companies are venture funded just like the American companies," Lee said. "They just want to serve their investors."

Beijing's role in the sector remains substantial, even under Lee's market-driven framing. Investors with ties to the government participated in more than 140 Chinese AI deals in 2025, Fortune reported earlier this year, citing a Pitchbook analyst note. President Xi Jinping has also championed the industry. In 2025, Xi called for China to develop "an independent, controllable, and collaborative artificial intelligence basic software and hardware system."

Lee added that American companies have sound reasons to spend more than their Chinese rivals. Heavy investment attracts the world's best researchers and gives them enough computing power to experiment.

Myth three: Export controls explain China's frugal playbook

The third misconception, Lee argued, is that U.S. export controls alone pushed Chinese companies toward models that require less computing power, while American firms pour billions into data centers. The truth, he said, lies in different priorities.

"The Chinese companies are naturally parsimonious; not just the startups," Lee said. He added that "prudence and survival instinct" matter more to Chinese firms than the pursuit of AGI.

That frugality has a hard commercial edge. Because of their resourcefulness, Chinese companies can develop and operate models for as little as one-tenth the cost of their American competitors, Lee said. That price advantage makes Chinese AI products attractive to businesses unwilling to pay premium rates for AI services.

How large is the remaining gap?

The competitive picture still favors the United States on several measures. American companies produce more top-tier models and higher-impact patents than Chinese developers, according to a Stanford University study. OpenAI and Anthropic remain the world's most valuable AI companies.

OpenAI was valued at $1.25 trillion in a November funding round and is eyeing a potential IPO in 2027. Anthropic is preparing to go public as early as next week, seeking a valuation of up to $1.4 trillion in what could be one of the largest IPOs in history.

Yet DeepSeek's 2025 breakthrough reshaped sentiment inside China's AI sector, Lee said. Before DeepSeek demonstrated its model, Chinese developers believed their country lagged years behind the United States.

"It's a national pride thing, so people say, 'Okay, we thought we were never going to catch up. Now we have a shot.' So, it was quite a special moment," Lee said.