Original text: Delphi Labs Cofounder, @ZeMariaMacedo

Compiled by: Big Pincer | PANews Lobster

I flew to China with high expectations—thinking I would see seriously underestimated geniuses changing the world at a fraction of the cost. Two weeks later, I returned with a more complex answer:

Chinese hardware is quietly winning a war that Westerners are even unaware has begun; but in the software sector, valuation bubbles and homogeneous founder profiles are turning what should have been an explosive ecosystem into yet another game of hot potato.

I spent two weeks in China, visiting founders, venture capitalists, and CEOs of listed companies in the AI ecosystem. Before entering, I was optimistic about this ecosystem, hoping to find world-class AI talents making strides at valuations far lower than in the West.

When I left, my views had become more nuanced: I was more optimistic about hardware than I expected, but more pessimistic about software, and I also formed some surprising views about Chinese founders that astonished even myself.

The Founder Issue

The outstanding founders I have invested in share a common trait: independent thinking, rebellious spirit, high focus, and near-paranoia. They do not go with the flow, constantly questioning "why", rejecting second-hand wisdom. Their decisions seem baffling to outsiders but are utterly reasonable to them. There is a strong drive within them that is deeply rooted, often manifested as a life experience filled with obsession and excellence. Their life trajectories have a sense of "edginess," easily recognizable among the high-IQ crowd a VC interacts with.

Many of the Chinese founders I encountered are of another type—which surprised me.

They are exceptionally talented—coming from top universities, having worked at ByteDance or DJI, published papers in Nature, holding multiple patents. Achievements that only top technical talent in the West could possess are merely entry tickets here. Their work ethic almost surpasses anyone I have ever met. We hold meetings at any time, on weekends, and across cities. One founder even came to meet us on the day his wife gave birth!

However, independent thinking, rebellious spirit, and a vision from zero to one—these are hard to find. The backgrounds of different founders are highly similar, their business plans are more conservative and steady, and their ideas are often advanced versions (V2) of what already exists, rather than truly original bets. Given that China has cultivated such a large pool of technical talent, I expected to meet more people proposing ideas I had never heard of.

My judgement is: The Chinese education system can cultivate excellence, but it does not allow enough space for "deviation." It produces founders who can execute known problems to perfection, rather than those who present problems that "no one has ever realized are problems."

Venture Capital Reinforcing This Model

More interestingly, local investors are actively amplifying this trend.

Most Chinese funds base their entire investment logic on investing in the best alums from ByteDance or DJI—favoring prestigious backgrounds over individuality, valuing qualifications over judgment. The composition of VCs mirrors this: most come from big companies, consulting, or investment banking backgrounds, similar to European VCs a decade ago.

Ironically, the best Chinese founders in history—those who truly established generational businesses—never stayed in big companies at all. Jack Ma is a two-time college entrance examination failure and former English teacher; Ren Zhengfei founded Huawei after retiring from the military at 43; Liu Qiangdong started from a street vendor to create JD; Wang Xing dropped out of his PhD program and began from scratch in entrepreneurship. The recent Liang Wenfeng founded DeepSeek without ever having been employed anywhere, having only worked at his own company. These individuals are outliers, those without impressive resumes—exactly the type of people the current system tends to overlook.

Finding these individuals is where true excess returns lie, yet from my perspective, almost no one is looking in that direction now.

Shenzhen and the Hardware Ecosystem

In China, what amazed me most was not any startup pitch.

But rather Shenzhen's "hardware underground world"—in those workshops, engineers systematically acquire high-end Western products, disassemble them part by part, and reverse engineer them with meticulous precision. When I left, I truly was unsure whether most Western hardware founders really understood what they are competing against. The network effects here are not theoretical but real, inseparable, and built over decades.

The entrepreneurs we met substantiated this: Over 70% of hardware raw materials come from the Greater Bay Area, nearly 100% from China—enabling iteration cycles far beyond the capabilities of Western hardware companies.

Most founders I met are following DJI's model: doing consumer-grade hardware in a niche market—electric wheelchairs, mowing robots, next-generation fitness equipment—scaling revenues to eight or nine figures, then expanding into adjacent categories using their customer base or underlying technology. Some of these companies' scales have exceeded your imagination. The company that impressed me most is Bambu, a 3D printing company that most Westerners have never heard of, reportedly making $500 million in annual profits and doubling every year.

Pessimistic Outlook on Chinese Software

Upon leaving, I had more doubts about opportunities in China's software sector than when I arrived.

At the model level, China's open-source achievements are indeed impressive—but the closed-source models still show significant gaps compared to the strongest models in the West, and these gaps are likely to continue widening. The chasm in capital expenditure is enormous, GPU acquisition remains constrained, and Western labs are intensifying restrictions on model distillation. Revenue data already illustrates the problem: it was reported that Anthropic made $6 billion in just one month in February. The best ARR of Chinese models is still at the tens of millions of dollars scale.

At the startup level in software, the mainstream profile consists of former product managers and researchers from ByteDance, building agents or environment perception consumer software for the Western market.

While the talents are real, most of these products fall within the functional range that large labs would natively integrate—risking being rendered obsolete by an individual product release at any time. I am also profoundly shocked by the lack of large-scale, rapidly growing private software companies in China.

In the West, apart from model companies, many startups have achieved ARR in the nine to ten-digit range with astonishing growth rates—Cursor, Loveable, ElevenLabs, Harvey, Glean. This type of breakthrough private software company essentially does not exist in China—while a few exceptions, such as HeyGen, Manus, and GenSpark, once they found breakthroughs, ultimately chose to leave.

Valuation Bubble

Despite the poor situation in the software sector, bubbles are very real—both in early and late stages.

In the early stage, although the costs of top talents from ByteDance, DeepSeek, and Dark Moon are still significantly lower than their American counterparts, the median valuations have converged. It has become common for consumer-facing startups pre-product to be valued at $100 to $200 million, and seed rounds exceeding $30 million are not surprising at all.

In late stages, the figures become even less sustainable. MiniMax's valuation in the open market is about $40 billion, while its ARR is under $100 million—approximately a 400x price-to-sales ratio. Zhitu AI is valued at approximately $25 billion, with revenues around $50 million. In contrast, OpenAI's peak valuation was about 66 times its ARR, and Anthropic was about 61 times.

Private model companies like Moonshot are using these public companies as benchmarks, raising funds at valuations of $6 billion, $10 billion, and $18 billion within just a few months. Those in the crypto circle should be familiar with this dynamic—investors are comparing private valuations with IPO stock prices before unlocking.

Furthermore, part of the reason supporting Zhitu and MiniMax in maintaining current valuations is that they are currently the only channels for exposure to the Chinese AI narrative, which carries a premium in itself. However, as more companies go public and this scarcity is diluted, the situation will change. Lastly, the IPO window often closes quickly and without warning—before you complete arbitrage, the benchmarks you referred to may have already changed, and that is no guarantee.

The humanoid robot sector faces a similar situation. There are approximately 200 humanoid robot companies in China, about 20 of which have raised more than $100 million, several with valuations in the billions—all of which are at the revenue pre-stage, most planning to list on the Hong Kong Stock Exchange in 2026 or 2027. If this market is real, China's hardware dominance provides a clear long-term trajectory. However, the commercial rollout is likely to be much slower than the current pace of financing implies; I am deeply skeptical about whether the Hong Kong market can accommodate numerous billion-dollar humanoid robot companies currently in the queue. I choose to observe for now.

Noteworthy Asymmetry

One thing I did not anticipate: almost every founder I encountered is prioritizing global markets over the Chinese market. They use Claude Code, follow Dwarkesh, and are well-versed in the startup ecosystem in San Francisco, often knowing more than Western investors who have not paid close attention.

The hostility from the West towards China far exceeds China's hostility towards the West. Chinese founders see no contradiction in combining China's engineering execution with hardware depth and the West's market expansion capability and product vision. When this combination becomes a reality in the right founding teams, it will give birth to some truly remarkable companies.

Finding those founders—those who do not fit the "credential template" that local VCs optimize for—is precisely the direction we are focusing on.

Special thanks to@woutergort for opening his amazing network in China to us, and thanks to@PonderingDurian

for organizing this trip, and thank you Claude for patiently helping me organize my fragmented thoughts on the plane.

Note: This article was mainly compiled by AI.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。