A Western VC's two-week observation of Chinese AI: hardware shocks, software pessimism, founders surprised

Author: José Maria Macedo, Co-founder of Delphi Labs; Compiler: Jinse Finance Claw

I spent two weeks in China, meeting founders, venture capitalists (VCs), and CEOs of publicly listed companies across the entire AI ecosystem. Before I went, I was optimistic about this ecosystem, expecting to find world-class AI talent starting companies for a fraction of Western valuations.

By the time I left, my perspective had become more nuanced: confidence in hardware was stronger than I anticipated, while I was more pessimistic about software, and certain views on Chinese founders surprised me.

Founder Issues

The great founders I have invested in all share highly recognizable traits: independent thinking, rebelliousness, extreme focus, and near-paranoia. They do not follow orders, constantly ask “why,” and refuse to accept conventional wisdom. Their decisions may seem incomprehensible to outsiders, but to them, they are self-evident. They possess an inner strength that is relentless, often manifesting as a long-term obsession with excellence. Their lives have a certain “edginess” that allows you to identify them among countless high-IQ talents seen by VCs.

However, many of the Chinese founders I encountered belong to a different archetype—this surprised me.

They are exceptionally talented—graduates from prestigious universities, having worked at ByteDance or DJI, published papers in Nature, and hold multiple patents. Achievements that are the hallmark of only the top technical geniuses in the West are merely “admission tickets” here. They are also among the most diligent people I have met. We held meetings at any time, on weekends, and in different cities. One founder even came to meet us on the day his wife was giving birth!

Yet, independent thinking, a rebellious spirit, and the vision to go from 0 to 1 were hard to find. The founders’ backgrounds are highly similar, their business plans tend more toward risk aversion, and their ideas are often “impressive V2 versions” of existing things rather than truly original bets. Given the scale of technical talent produced in China, I had expected to see more unheard-of brilliant ideas.

My view is that China’s education system produces excellence but does not leave enough space for “deviating from the norm.” The result is that these founders are genius executors solving known problems rather than those who can identify “problems that no one knows exist.”

VCs are Reinforcing This Model

Interestingly, local investors are actively exacerbating this model.

The investment logic of most Chinese funds revolves entirely around supporting elite alumni from ByteDance or DJI—valuing background over edginess, and credentials over conviction. The background of VCs themselves reflects this: most come from large companies, consulting, or investment banking, similar to European VCs a decade ago.

Ironically, historically, the best founders in China—those who truly built epoch-making companies—never worked in large corporations. Jack Ma was an English teacher and failed the college entrance exam twice; Ren Zhengfei founded Huawei after leaving the military at 43; Liu Qiangdong started by selling goods in Zhongguancun; Dr. Wang Xing dropped out and began his entrepreneurial journey from day one. A recent example is Liang Wenfeng, who had never worked elsewhere before founding DeepSeek. These are “outliers,” lacking the so-called credentials—precisely the type of profiles that the current system tends to overlook.

There are real excess returns (Alpha) in finding such profiles, but currently, it seems few are paying attention.

Shenzhen and the Hardware Ecosystem

What shocked me most in China was not the startup pitch events.

It was the underground hardware world in Shenzhen—where engineers systematically obtain high-end Western products, disassemble components one by one, and carry out reverse engineering with extremely tight precision. By the time I left, I was genuinely unsure whether most Western hardware founders understood what kind of competitors they were up against. The network effects here are not theoretical; they are physical, dense, and have been built over decades.

The entrepreneurs we met validated this with data: over 70% of hardware investment comes from the Greater Bay Area, nearly 100% from within China—this allows for iteration cycles that Western hardware companies simply cannot match.

Most founders I encountered follow the “DJI model”: creating consumer-grade hardware in a niche market (such as electric wheelchairs, weeding robots, new generation fitness equipment), scaling to 8 to 9 figures in revenue, and then leveraging their customer base or underlying technology to expand into adjacent categories. Some of these businesses have already scaled beyond imagination. The most impressive company I met was Tuozhu, a 3D printing company that most Westerners have not heard of, reportedly achieving an annual profit of $500 million and doubling every year.

Pessimism About Chinese Software

By the time I left, my skepticism about opportunities in Chinese software had exceeded my initial arrival doubts.

At the model level, China’s open-source models are indeed impressive—but closed-source models significantly lag behind the best in the West, and the gap may widen. There is a massive disparity in capital expenditures (CapEx), and GPU acquisition remains constrained. Western labs are increasingly tightening restrictions on model distillation. Revenue data clearly illustrates the problem: reports indicate that Anthropic made $6 billion just in February. Meanwhile, the annual recurring revenue (ARR) of the best models in China is only in the tens of millions.

In software entrepreneurship, the mainstream profile is former product managers and researchers from ByteDance, building agent or environment perception consumer software targeting Western markets. The talent is indeed present, but many products fall within the scope of the native functionalities of large companies—once a large company releases a version, these products become redundant. I was also shocked by the lack of large, fast-growing private software companies. In the West, aside from model companies, several startups (like Cursor, Loveable, ElevenLabs, Harvey, Glean) are achieving astonishing growth rates, printing 9-figure and 10-figure ARR. Yet, such breakout private software companies at this level barely exist in China—rare exceptions like HeyGen, Manus, and GenSpark, once they spot opportunities, ultimately choose to leave.

Valuation Bubble

Despite the bleak prospects for software, there is a real bubble in both early and late stages.

In the early stage, while top talent from ByteDance, DeepSeek, and Moonshot is still significantly cheaper than their U.S. counterparts, median valuations have converged. It is quite common to see consumer startups valued at $100 million to $200 million before they even release a product. Seed financing exceeding $30 million is also not unusual.

In the late stage, the numbers are harder to justify. Minimax is trading at around $40 billion in the public market, with an ARR of less than $100 million—about a 400x sales multiple. Zhizhu is valued at around $25 billion with $50 million in revenue. In contrast, OpenAI’s peak fundraising round had a sales multiple of about 66x, and Anthropic about 61x.

Non-public companies like Moonshot are leveraging these publicly listed benchmarks to raise funds at valuations of $6 billion, $10 billion, and $18 billion within a matter of months. Cryptocurrency investors would recognize this dynamic: investors are comparing private valuations against yet-to-be-unlocked fair prices. Additionally, part of the reason supporting Zhizhu and Minimax at these valuation levels is that they are currently the only means to gain exposure to the “China AI narrative,” thus carrying a premium. This situation will change as more companies go public and dilute this characteristic. Finally, the IPO window often closes very quickly and without warning—you cannot be sure to complete such arbitrage before benchmark prices shift.

The humanoid robot sector is similar. China has about 200 humanoid robot companies, with around 20 raising over $100 million, and a few reaching billion-dollar valuations—almost all have no revenue, with most planning to IPO in Hong Kong in 2026 or 2027. If this market is real, China’s dominance in hardware makes long-term outcomes quite clear. However, the commercialization may be slower than the current funding pace suggests, and I question whether the Hong Kong market can support so many humanoid robot companies planning to raise billions. I am currently choosing to wait and see.

Asymmetries Worth Noting

One thing I did not anticipate: almost every founder I met is building products for the global market, not just for the Chinese market. They use Claude Code and watch Dwarkesh’s videos. They are well aware of the San Francisco startup landscape, often more so than Western investors who have not paid much attention to the details.

Western hostility toward China runs much deeper than Chinese hostility toward the West. Chinese founders believe that combining China’s engineering execution capabilities and hardware depth with Western go-to-market (GTM) strategies and product visions is not contradictory. This combination, when present in the right founding teams, will yield some truly exceptional companies.

Finding these founders—those who do not fit the “elite credentials” template optimized by the local VC ecosystem—is precisely where our current focus lies.