This Middle School Dropout Built His AI Startup In China. Now He’s Scaling It In Silicon Valley.

Luyu Zhang moved from China to the US last year and can barely speak English. He’s not trying to improve right now, either. That will come later.

“I’m too busy with work right now to improve my English,” he tells Forbes through a translator, sitting in a temporary office in Menlo Park. “That requires time, and right now I don’t have a day to lose.”

Zhang is the CEO of artificial intelligence startup Dify. He’s part of a growing cohort of Chinese AI founders uprooting their families and companies to move to Silicon Valley. On the surface, the trend seems counterintuitive. Washington has tightened export controls on advanced AI chips. Lawmakers have for years warned of IP theft and China’s military and economic ambitions. There is endless talk of decoupling in both countries.

And yet the founders keep coming.

Zhang’s bet is simple: if you want to build a global AI infrastructure company, you build where the world is competing.

A middle-school dropout and coding prodigy, Zhang recently raised $30 million at a $180 million valuation from HSG (formerly Sequoia Capital China) and other Asian venture capital funds Hillhouse Capital, 5Y Capital and Mizuho Leaguer Investment. Before starting Dify, he worked at multiple startups and later managed a large engineering team at Tencent.

Dify began as an open-source project that helps developers build AI applications using a low code interface instead of writing extensive backend code. It now ranks 52rd among the most-starred repositories on Github. The company built around it employs 100 people, is profitable and serves more than 280 enterprise customers, including Volvo, ThermoFisher Scientific and Novartis.

Zhang points to Meta’s proposed $2 billion acquisition of Manus—an AI startup founded in China before relocating to Singapore—as a blueprint. He calls it the “China origin + overseas operation” startup model. (Manus has faced scrutiny on both sides: U.S. regulators examined an investment made by U.S. venture capital firm Benchmark, and Chinese regulators are reportedly investigating whether to block Meta’s acquisition.)

Zhang says building a truly global company from within China is difficult. The country’s tech ecosystem is largely self-contained, and software developed there typically serves only the domestic market.

“If you’re an elite athlete, you don’t play only to win local matches. You train to compete in the Olympics. For a technology company building infrastructure for the AI era, Silicon Valley is our Olympic arena,” he says. “We want to compete at the highest level.”

Zhang says he knows nearly 20 Chinese startup founders now in the process of relocating to the US.

Investor Lake Dai, formerly head of product at Alibaba and now founder of Sancus Ventures, says the relocation trend is broader. In the past two years, she says, at least 100 Chinese entrepreneurs have contacted her about moving to the states. Foreign capital has begun to pull back from China, tightening venture funding and forcing founders to consider other markets. (Dai is not an investor in Dify).

“Chinese founders are moving to the US now,” Dai says. “We’ve seen more and more of that in the last few years.”

But nobody wants to end up like TikTok, which after years of scrutiny over its Chinese ownership, briefly went dark in the U.S., and was ultimately forced to sell its U.S. operations to avoid a ban.

To avoid backlash, many startups emphasize what their products do—not where their founders were born. “They don’t want to be labeled prematurely,” Dai said.

Growing strategically is a balancing act that Chinese-born startups are still figuring out. Dify does not hide its Chinese roots, and though Zhang says he’s hiring aggressively in the Bay Area and Tokyo, the company’s core open-source engineering team of 60 is still in China. Other startups have adopted similar hybrid structures, maintaining a core engineering team in China while hiring for functions like customer service and sales abroad. Some, like OpusClip or HeyGen, have fully relocated outside of China.

Zhang grew up in Anhui province. By middle school, he says he was earning about $1,000 a month coding websites—more than his father, a civil servant. He dropped out, unable to tolerate China’s famously rigid schooling system. His financial independence made it impossible for his parents to force him back. By 2018, he’d landed a product leadership gig at Tencent. After encountering generative AI in 2022, he started Dify to make it easier for developers to build and deploy AI applications at scale. A trip to Nvidia’s annual GTC conference convinced him to move.

“The energy was different,” he says.

Still, there are hurdles to overcome. U.S. policymakers warn about sensitive technology transfer to China’s AI sector. Some American investors avoid startups with Chinese backers. Meanwhile, anti-Chinese sentiment simmers in parts of Silicon Valley. Recently, a Chinese Anthropic AI researcher quit, saying on his blog that the company’s “anti-China statements” were a key reason.

When it comes to AI and Chinese entrepreneurs, Zhang argues the rhetoric needs to be more nuanced. Yes, we need to keep national security top of mind, he says, but not all AI products carry the same risks. Dify is an open source product that customers typically host on their own servers, placing it outside what he describes as a “sensitive sector”. It is different, he argues, from a consumer social platform with algorithmic influence or from companies developing export-controlled frontier chips.

“People go to CES and don’t have a problem buying Chinese-made consumer electronics,” he says.

And it’s not like there isn’t a place for them here. American startups have always relied on Chinese AI talent. A December study from the Carnegie Endowment tracked 100 top Chinese AI researchers working at US institutions and companies in 2019. As of December, 87 were still in the U.S.; ten had returned to China. Of Meta Superintelligence Labs’ entirely-immigrant founding team of 11, seven were born in China.

So while Washington and Beijing may be drifting further apart, in AI at least, talent isn’t decoupling. Dai says we need to move past blanket suspicion, while maintaining smart scrutiny.

“I know people are concerned and saying, what are these Chinese citizens doing here in the AI industry?,” she said. “But I think we should figure out a way to welcome these amazing founders here. They want to build the next generation of AI companies here and that benefits all of us.”

For founders like Zhang, the calculation is less ideological and more practical: if Silicon Valley is the deepest pool of AI talent, capital and ambition, that’s where he wants to be. He says that he doesn’t want to take sides in an AI race; he just wants to compete in the arena.

“Entrepreneurs coming from China to the US are not political,” he says. “We just want to build great products that people use.”

More from Forbes

ForbesThe 50 Hottest Fintech Startups In 2026By Jeff KauflinForbesInside Phantom’s Battle Plan To Win The Crypto-Fueled SuperApp WarBy Nina BambyshevaForbesMany Countries Pay Big Bonuses For Olympic Medals. This One Is On The Hook For $7.8 Million.By Brett KnightForbesThis Fintech Is Winning With Hispanic Immigrants. Now It Has To Survive Trump.By Sophia Acevedo