The investor who backed the "Four Kings" of computing power wants to stay quiet amid the embodied intelligence boom.

90 billion yuan of hot money has flooded into the embodied intelligence track, and the battle for the "20 billion valuation club" has reached white-hot intensity. Tickets for WAIC (World Artificial Intelligence Conference) are almost impossible to get, and the number of exhibitors in the embodied zone has skyrocketed from 80 to more than 200... In these six months when the whole nation is fixated on robots, Wang Guangxi chose to "step back".

"When WAIC is packed with so many people this year, that's exactly the time we shouldn't show up." At the end of July, he told *China Entrepreneur*, "VCs that invest in cutting-edge and early-stage technologies should go to the scene when no one else is paying attention."

Wang Guangxi works at Lenovo Capital, one of the institutions that has made the deepest layout and reaped the most benefits in this boom. Covering four leading computing power players including Hygon Information, Cambricon, Moore Threads and MetaX, as well as embodied intelligence stars like Star Motion Era and Motion Dynamics, the portfolio of this veteran CVC has almost covered all key links in the AI embodied intelligence field. While others are still crowding outside to watch the bustle, he has already laid his cards on the table.

But this is not a success story about "how to bet on the right track". It is a sample of how an investor maintains independent judgment amid noisy consensus, and fights against human nature as well as his own prejudices. We hope you can gain the following three insights through this interview:

First, understand the real temperature of the embodied intelligence track. Is there a bubble? Why are people still trapped in cutthroat competition? Wang Guangxi's judgment is: there is a bubble, but the key is not to be carried away by emotions. He believes that next year will bring a survival test, and IPO will be the turning point. For entrepreneurs and investors who want to see the direction of the track clearly, this is a worthwhile reference timetable.

Second, learn about the investment logic of a CVC operator. From gathering the "four giants" of computing power to betting on embodied unicorns, what is Wang Guangxi's sense of rhythm? Why can he place bets in advance? How to distinguish companies worth accompanying for 10 years from "firework projects" that fade quickly? How to respond to the doubts about "scattershot investment"? "Many people spread their bets wider than us, and are luckier to hit more targets than us. We are treading on thin ice all the time." This sentence may give you a glimpse of the answer.

Third, think about the changes in human choices and relationships in the AI era. The generational shift between two generations of AI entrepreneurs is taking place. From AI 1.0 to AI 2.0, the biggest variable is original aspiration and motivation. What is the most common pitfall young entrepreneurs fall into? How can senior investors be changed by young people? When the final outcome that AI replaces human beings comes, what value will be left for humans? There are no standard answers to these questions, but Wang Guangxi's response provides a unique perspective.

"In the thousands of years of carbon-based narrative of human history, we have always had a heart pursuing truth. If silicon-based intelligence can really replace carbon-based life, how can you be sure it will not pursue truth even more persistently?" In his opinion, that does not necessarily mean degradation, but evolution. Interestingly, Wang Guangxi told us that he is not worried about being replaced by AI for the time being, because "investment is a people business, and most of the key things are not written on paper".

Wang Guangxi, Ph.D. in Electrical Engineering from Caltech, is currently Vice President of Lenovo Group and Managing Partner of Lenovo Capital. He joined Lenovo in 2011, and became part of Lenovo Capital when it was founded in 2016. He is an introvert, teases himself as a "straight engineering guy", prefers quietness, has concept-driven thinking, likes to break down problems, and places bets only after thinking things through thoroughly. He never shuns past misjudgments, and even when he says some questions "are hard to have answers", he will give his judgment every time.

In his view, quiet thinking is also an important capability for VC investors.

At the end of the interview, we talked about *The Matrix*, an old movie from the 1990s. Facing the red and blue pills, Wang Guangxi made the same choice as the protagonist in the movie, preferring to face the cruellest and most real state directly. Once people have sunk costs or specific demands, it is easy for them to deceive themselves. But he chooses to keep fighting against himself. "You have to keep exploring the unknown, getting closer to the truth, even if that truth makes you uncomfortable."

Photography: Deng Pan

The following is the exclusive interview transcript of Wang Guangxi by *China Entrepreneur* (abridged):

01

On "90 Billion Hot Money and Bubble": The Real Temperature of the Embodied Intelligence Track

*China Entrepreneur*: In the first half of this year, everyone was competing for the admission ticket to the "20 billion valuation club", with more than 90 billion yuan of capital chasing projects, but some predictions say that no more than 10 companies will survive in the end. Do you have FOMO (fear of missing out)?

Wang Guangxi: There are already more than enough companies with valuations of 10 to 20 billion yuan that I can count with two hands. There is definitely a bubble. So many companies have seen such huge valuation changes in such a short time, which is not in line with the laws of technology and business. A lot of the growth behind it is not supported by commercial or even technological breakthroughs.

We are in a good state, because we invested very early. We have been paying attention to the entire robot track since 2017 and 2018, and have invested in more than 40 companies so far. For this embodied intelligence boom, we started researching and placing bets back in 2023, so we don't have much FOMO. Otherwise, watching a high-speed train start up and deciding whether to cling to it and climb on is really not a good experience for investment.

Source: AI generated

*China Entrepreneur*: But you made intensive layouts as early as 2023, half a year earlier than the market. Isn't that an advance version of FOMO?

Wang Guangxi: We saw the changes brought by GPT, and realized that intelligence would extend from the digital domain to the physical domain, so we started placing bets. What does FOMO mean? You rush to grab things just because you see everyone else is grabbing. When we started our layout, no one else was even paying attention to this track.

*China Entrepreneur*: Embodied intelligence is not like the internet where the winner takes all, so why are people still competing for rankings and the title of "number one"?

Wang Guangxi: It is really hard for the winner to take all. The internet has a very clear traffic aggregation effect, but it is already a bit difficult for the AI sector. Especially when it comes to physical intelligence, which involves a large number of software and hardware solutions combined with specific scenarios, it is even harder for the winner to take all.

*China Entrepreneur*: Then what are people competing so fiercely for?

Wang Guangxi: Many entrepreneurs in this fast-changing and highly competitive environment, especially in China, need to work on technology, business, financing, PR, and ranking campaigns at the same time. I can understand that from the perspective of competitive pressure. But indeed, many companies have completed two or three consecutive rounds of financing, even within one or two months or at the same time, with price differences of several times, while their fundamentals have not changed at all.

*China Entrepreneur*: As an investor, what do you firmly believe in?

Wang Guangxi: We have always adhered to a very clear belief that this industry will eventually generate huge value. We will not invest just because the market is hot now and many big names are entering the market, nor will we invest just because many other institutions are investing, nor will we invest just because the company can go public or its valuation is hyped very high. We really have few distracting thoughts in this regard.

02

On "Greatness and Fireworks": What Kind of Company Is Worth Accompanying for 10 Years?

*China Entrepreneur*: What kind of company is worth accompanying you for 10 years?

Wang Guangxi: In this wave of AI and embodied intelligence track, especially some companies with grand blueprints that aim to build platform capabilities, are all worth accompanying for more than 10 years. To build a generally recognized great company, it really takes a timeline counted in decades.

*China Entrepreneur*: How do you define greatness?

Wang Guangxi: One of the positioning points of Lenovo Capital is that we want to invest in great companies of the next era. The meaning is very simple: we hope that this company can truly use technological innovation to bring breakthroughs in products and services, truly serve the society, and create huge value for it.

*China Entrepreneur*: That sounds a bit empty.

Wang Guangxi: It's easy to understand. Today when you open your phone, you can see so many things you use every day, but calm down and think: 10 or 15 years ago, these things might not have existed at all. What was your life like back then? It was completely different from today. That is greatness.

*China Entrepreneur*: What about those companies that only make dancing robots? Are they great or just fireworks?

Wang Guangxi: None of those companies that make performance and dancing robots will position themselves as "I am just a dancing robot company". I also believe no company will tell investors or the public that "I will only make dancing robots in the future". Dancing is just a demonstration of staged motor capability. Those companies that currently seem to use performance and electromechanical capability demonstration as their main PR means are actually platform-oriented companies, and their ambitions go far beyond making dancing robots.

However, I predict that by next year, startups whose technology can't keep up, or who still stay in the demo stage and can't achieve commercial landing, will face a survival test. The IPO of embodied intelligence companies will be a turning point. After going public, enterprises become public companies under the spotlight, and will undergo the "interrogation" of the secondary market. Their operating conditions such as financial information will be more open and transparent, but the companies will also face greater pressure as a result.

03

On "Investment Rhythm": "You Should Go to the Scene When No One Else Pays Attention"

*China Entrepreneur*: There is a view that AI is moving from the narrative stage to the delivery stage. When will this turning point happen specifically?

Wang Guangxi: From the social perspective, I don't think it has happened yet. You can see a lot of people arguing under every post about embodied intelligence. But in the investment circle, no matter because of FOMO or other reasons, everyone's consensus on this matter is indeed a bit too high.

*China Entrepreneur*: Isn't high consensus a good thing?

Wang Guangxi: High consensus itself is not a problem, but the pressure brought by excessively high consensus is that valuations rise ridiculously high, and the rhythm also becomes abnormal.

*China Entrepreneur*: Now when you go to inspect projects, what can tell you "this company is not bragging"?

Wang Guangxi: The embodied intelligence field is extremely comprehensive, covering hardware, software, interaction, production, supply chain, quality control, and industry know-how, all of which need to be mastered. You can't judge just by checking "whether this technical point is real or not". If the company's core business is R&D and intelligence, then the technical foundation of the team itself and their ability to catch up with the cutting edge quickly are particularly important.

Source: AI generated

You can't expect one company to handle all the important things in the industry on its own.

*China Entrepreneur*: You started paying attention to embodied intelligence in 2022, but you didn't make any moves for half a year. Then you made a batch of intensive investments in 2023 and 2024, and the rhythm slowed down again after 2025. Between the acceleration and deceleration, what did you observe?

Wang Guangxi: We started to speed up in mid-2023, and the most intensive period of investment was 2023 and 2024.

By 2025 and this year, we have basically completed the layout for the first echelon or the leading complete machine solution enterprises. Because we made moves relatively early, we didn't invest too much money in the very early rounds.

In the next one or two years, we will focus more on finding new variables, such as upstream core components, companies with special resources and landing capabilities, or very early startups with different technical routes. Essentially, it is continuous tracking and filling gaps in our portfolio.

*China Entrepreneur*: At this year's WAIC, the embodied exhibition area expanded from 80 exhibitors to more than 200, and everyone was crowding to watch. Did you go?

Wang Guangxi: I didn't go this year, but I went in previous years. When there are so many people this year, that's exactly the time we shouldn't go.

VCs that invest in cutting-edge and early-stage technologies should go to the scene when no one else is paying attention, attend academic conferences together with professors and students. When the whole nation rushes to WAIC, it's time for us to step back.

04

On "Brain Competition": Routes Not Converged, Essence Clear

*China Entrepreneur*: The embodied "brain" now has three parallel routes: the layered school, the unified school, and the scenario school, and no one has convinced anyone else. Do you have the answer?

Wang Guangxi: When we first invested in embodied intelligence, we skipped those companies that only make the brain. At that time, we judged that under the domestic business environment, the route convergence is far away, and the business model of only making the brain is very difficult, because the brain cannot directly serve customers, and those companies that make robot bodies are also developing their own brains, which may be in a cooperative and competitive relationship with you. Looking back, several brain-focused companies we evaluated back then have all started to make hardware themselves in the end.

But the brain itself is definitely important. The core reason why embodied intelligence is so popular this time lies in intelligence. Robots have existed for decades. In the past, people thought they were just mechanical arms, an extension of automation. But the biggest variable this time is the brain.

*China Entrepreneur*: Which route do you think will win?

Wang Guangxi: Now the route framework has not converged, all companies are evolving, and there will be staged differentiation according to their own technical strengths and the scenarios they want to land. In the long run, technology will gradually converge.

*China Entrepreneur*: When do you think the GPT