Linear Capital Interviews CHEN Yilun of Tashizhihang: In the Trillion-Dollar Embodied AI Arena, Only the Victor is the Hero

"Between the Lines" is an in-depth dialogue video podcast produced by Linear Capital. The guest of our second episode is Chen Yilun, Founder and CEO of TARS Robotics.

One year ago, Chen Yilun left Tsinghua AIR (Institute for AI Industry Research) and co-founded TARS Robotics together with Li Zhenyu, former President of Baidu Intelligent Driving Group, and Ding Wenchao, former Huawei "Genius Youth", among others. The company raised $242 million in its angel round, setting the record for the largest angel financing for embodied intelligence in China, and Linear has supported the company for three consecutive rounds. Contrary to the company's popularity in the capital market, the TARS team lives up to its name, being down-to-earth and low-key, and has rarely shared their founding story publicly.

Most people only see the speed of this "rocket", but this time we want to talk about the person steering this rocket. This is Chen Yilun's first appearance on a video podcast since founding TARS, and the first full presentation of the company's founding story. We hope it can satisfy some of your curiosity about this company.

TARS Robotics is probably the most "counter-mainstream" player in the embodied intelligence track — it set the $242 million angel round record when the industry was still lukewarm, talked about World Engine when VLA was the hottest trend, and insisted on a human-centric approach when teleoperation was prevalent.

Contrary to its market hype, the TARS team lives up to its name, being down-to-earth and low-key. Most people only see the speed of this rocket, with equal parts curiosity and doubt. So today, we want to talk about the person steering this rocket — Chen Yilun, Founder and CEO of TARS Robotics.

Every major choice Chen Yilun made at key life moments seems confusing at first glance: He got a guaranteed admission to Tsinghua via the National Physics Olympiad, and chose the Department of Electronic Engineering. After graduating with his PhD, he turned down a $500,000 annual salary offer to learn hydraulics at an electromechanical company. He was happy working at DJI, but left anyway. He built Huawei's intelligent driving business from scratch and made it an industry sensation, then turned around and returned to Tsinghua to conduct research.

He says he has always loved "things that move", and wants to find a truly meaningful big problem in the robotics field.

In February 2025, TARS Robotics was founded, building an all-rounded founding dream team, closed three consecutive rounds of financing, and reached the top tier of the industry in just one year. But Chen Yilun says, Our business plan hasn't changed across all rounds, we've just been turning what was on paper into reality step by step. Instead of starting with simple tasks like box moving, TARS went straight for the hardest challenge in manufacturing: wire harness processing. In China alone, there are still 1 million industrial workers working in this field today.

In Chen Yilun's view, there are no born heroes that can win wars. Only when you gather talents together to win the war and complete the mission of the era, do these people become heroes.

Welcome to step into the story of Chen Yilun and TARS Robotics, in this booming trillion-dollar embodied intelligence industry. This is not only the first full presentation of TARS's founding story, but also an answer to how a "light chaser" chose to become a "spark".

Part 01 The Pivotal Moments That Shaped Fate

A Physics Kid's Wuxia Fantasy of "Internal Strength"

▍Harry Wang Huai: Yilun, we know your latest financing round went very smoothly, and we've supported you three times in a row. Although TARS's story has just begun, it has already been an extremely brilliant debut. Embodied intelligence as a field is also being recognized by more and more people.

They say heroes don't dwell on past glories, but everyone's present can more or less trace its roots back to the past. I'm curious: you got guaranteed admission to Tsinghua via the Physics Olympiad, but you majored in electronics. What was your thinking back then?

▍Chen Yilun: I grew up and attended middle school living with my grandparents. They didn't like watching TV, so there was nothing fun to do at home, only two types of books: a pile of physics books, and a pile of wuxia martial arts novels.

Gradually, I developed two big hobbies: First, I love reading wuxia novels, I've read almost every wuxia novel out there. There's a classic trope in wuxia stories: the protagonist falls off a cliff, accidentally finds a secret martial arts manual, then trains hard, his internal strength grows until he masters an invincible great skill. These stories always got me so excited. Second, I love physics extremely. When I studied physics hard, I would fantasize that one day I could also master great skills and become an excellent scientist.

▍Wang Huai: Physics and wuxia master, that's a really interesting combination.

▍Chen Yilun: When I was in middle school, I always thought I was very talented in physics. I finished all high school physics in junior high, I started learning college general physics in 8th grade, and finished differential equations by 9th grade. But all of this actually came from my immersion in that wuxia idea of "training hard to build internal strength".

Later I joined the Physics Olympiad, and I was determined to become a great physicist, maybe even someone like Einstein. But then a few things changed my perspective, and they have benefited me ever since.

The first thing was that I made it all the way to the national training camp for the Physics Olympiad smoothly, where 5 out of 25 students would be selected to represent China at the International Physics Olympiad. We trained in Shanghai, and after about a week, you realize a very sobering fact — everyone thought they were the next Einstein before they came, but once you're there, you find out all 25 of them think the same thing.

▍Wang Huai: But in the end 20 people have to leave disappointed.

▍Chen Yilun: So that's when I realized that at least I wasn't the one unique one, and I didn't end up making the top five anyway, but this experience taught me a lot.

The second thing was, I had two favorite books when I was in middle school. One was the Landau Course of Theoretical Physics, which explains the principle of optimization (all things operate following the minimum energy consumption). When I read it I had a real epiphany, it made me believe the world definitely has underlying laws. The other was the Feynman Lectures on Physics, he explains complex physical concepts in such a clear and approachable way. I found this whole process so interesting, that I realized I might be an engineer at heart. Later, about half of our training camp went to Peking University's Physics Department, and almost all of those who went to Tsinghua, without exception, went to Tsinghua's Department of Electronic Engineering.

▍Wang Huai: I remember Feynman said that, if you can't explain a concept clearly to an 8-year-old child, you probably don't really understand it yourself.

▍Chen Yilun: That's exactly right, this is what people later call the Feynman Technique, and I've tested it personally and it works.

Born to Love "Things That Move"

▍Wang Huai: You've said before you're born to love things that move. When did this start?

▍Chen Yilun: There's a big difference between physics and mathematics: when you solve a problem or do a derivation in math, a lot of the time you don't know what it's actually used for. But with physics, after you work through something, you can verify it in the real world, and that's so exciting. Back then, the provincial team sent me to a university to do extra physics experiments, and I went through all the university physics experiments. I felt so great back then, this was way more fun to me than deriving formulas.

▍Wang Huai: So you've been the type that "would rather do it than talk about it" since then. I remember another pivotal moment for you was in 2007, when you saw Boston Dynamics' robot dog on ice. Can you describe what that was like back then?

▍Chen Yilun: That's right, I was doing my PhD at the University of Michigan back then, my research direction was statistical machine learning, which is actually the predecessor of today's AI. Those years I almost became a mathematician, I spent every day in the library looking through all kinds of mathematical theorems, to prove that the method I proposed satisfied some optimality principle.

But the University of Michigan is actually very strong in robotics, the famous ostrich leg robot came from Professor Jessy W. Grizzle in our department. So you can imagine, at that time, on one side I was deriving formulas every day, and on the other side I saw my roommates working with moving robots in the lab, and I was extremely envious. But I also noticed that they were still using some very traditional methods, so even though the robots could move, they were very clumsy.

In 2007, Boston Dynamics already posted videos on YouTube. That hydraulic robot dog could slip on ice, get pushed, and still maintain balance — this was extremely shocking for anyone studying algorithms or robotics. Back in that era, this level of performance was generations ahead. And unlike academic work, they didn't release any papers for you to study how they did it. So this industrial work of theirs sparked my curiosity even more, it got me completely hooked.

▍Wang Huai: Sometimes people just get obsessed with hard things. How much did this affect your later career and startup choices?

▍Chen Yilun: It had a huge impact. When I saw that, I knew this was what I wanted to do. I had spent a lot of time on algorithms before, where you input a bunch of numbers and output a number, but what does that number even mean? But if you can directly perceive that number, even see it create value, the satisfaction you get is completely different.

Besides the Boston Dynamics robot dog, Tesla's electric car also gave me a huge shock back then. Especially after the Roadster, Tesla's first production model, came out, I thought it was so cool. These moving things made me realize how powerful good algorithms can be.

▍Wang Huai: It felt like seeing the light, right?

▍Chen Yilun: Exactly, I saw the light, so I had to go chase it.

▍Wang Huai: Looking back, there are always things that attract you and keep you up at night at many stages of life. Those are the lights that guide where you should go in life, you can't just let them slip away after one night.

Giving Up a $500k Annual Salary for His Dream Starting Point

▍Wang Huai: With the direction you studied for your PhD, it shouldn't have been hard to get a job at top Silicon Valley companies back then, but you chose to go learn hydraulics at an electromechanical systems company. How did you make that choice?

▍Chen Yilun: It really was a very counter-mainstream choice. When I graduated, my advisor strongly urged me to stay in academia, he said "Yilun, I think you are suited to be a professor". But I prefer doing practical things, so I ended up with three offers.

The first offer was to do quantitative trading on Wall Street, the annual salary was already $500,000 in 2011, that was definitely a huge amount of money. The second was more conventional, it was Google, dozens of people wanted to refer me back then. The salary was around $150,000 a year, which was already very good for a newly graduated PhD.

▍Wang Huai: That salary is really rare, it's the first time I've heard of dozens of people all trying to refer the same person — everyone just thought the referral bonus was easy money right? Haha.

▍Chen Yilun: The third one, which I chose, was a company called Eaton. I had never heard of this company before, but there was a senior alumnus from Tsinghua working there, he told me it was a Fortune 500 company, and the most important thing was that his department was the Innovation Center, which had a very interesting operating model.

The department was newly established, reported directly to Eaton's CTO, and got a lot of attention. And their way of working was very similar to a startup, the head came from Siemens TTP division, and they basically ran the department like a venture capital firm. Employees proposed ideas, and the company funded them through Stage A, B, C, to turn the idea into a product, even bring it to market.

▍Wang Huai: That's a really interesting rule set.

▍Chen Yilun: He told me, don't treat this as just a job, it can be a learning experience for you, it's like getting an MBA. And as an innovator, I could move around Eaton's different business units. I went to look at what businesses Eaton had, and I got more and more hooked, especially they had a hydraulics business, which reminded me of the Boston Dynamics robot dog. Back then, Eaton was ranked number one in hybrid powertrain for commercial vehicles.

I realized that I had always had a vague idea to do robotics, but I only knew algorithms, I didn't know anything else. Here I could work on these things like a startup, and collaborate with experts from various fields, it was perfect.

▍Wang Huai: This was the first door to the dream you were chasing?

▍Chen Yilun: Yes, Eaton was a really valuable experience for me.

▍Wang Huai: If you want to attract top talent, money alone is not enough. For people who are really talented and also have dreams, you have to win them over with dreams. How long did you stay at Eaton? How did you end up joining DJI?

▍Chen Yilun: I actually stayed at Eaton for quite a long time, around five years. We were a global innovation center, and they were going to build a center in Shanghai, so I was one of the first four people who came to build it from scratch.

▍Wang Huai: That was a startup experience in itself.

▍Chen Yilun: You could say it was completely from zero to one: we first defined what we wanted to do, then we converted an underground garage into a big lab ourselves, recruited a global team, it was really interesting.

Part 02 The Training at DJI and Huawei

The Most Important Lesson I Learned at DJI

▍Chen Yilun: There were two reasons I left Eaton. The objective reason was that the original CTO was an old Swedish gentleman who was very inclusive of innovation. After he retired, an Indian-American CTO took over, he drastically cut the budget, and a lot of things changed from how they were before.

And the subjective reason was, even though I was really into what I was doing, I realized that what I was doing was getting further and further away from machine learning and AI, and those years were exactly when AI was growing the fastest.

And I kept in touch with my classmates and friends, so I knew everything that was happening in Silicon Valley, I knew AI was progressing really really fast. I thought about it, and I still wanted to do robotics.

Where is the best robotics team in China? Most people would say DJI. Back then robotics wasn't a big industry yet, the best people in academia would just compete in contests, and people who wanted to work in robotics would go to DJI