Post-00s Huang Yi founded the full-stack open-source humanoid robot company RoboParty and launched the RP1

He is Huang Yi, founder of the humanoid robotics company RoboParty.

Huang Yi, born in 2004, is the founder and CEO of RoboParty. He founded a full-stack open-source bipedal humanoid robotics company, building an open-source ecosystem platform and constructing the R&D foundation for embodied intelligence.

Being a "pirate captain" on the boundless sea

Huang Yi's connection with robots began with a drone. In 2022, shortly after entering the School of Future Technology at Harbin Institute of Technology, Huang Yi went straight into the lab and led a team to develop a land-air amphibious drone, which won first prize in a national college student science and technology competition.

A year later, at the end of 2023, he spent less than 20,000 yuan in his dormitory to purely "hand-build" a bipedal robot that could walk, AlexBot. From mechanical structure design and joint modules to motion control algorithms, Huang Yi did it all himself, and he also open-sourced the entire R&D process, code, and more, publishing everything publicly on virtually every developer community. This series of moves made him famous overnight and quickly drew industry attention.

AlexBot, the open-source bipedal humanoid robot "hand-built" by Huang Yi

Z Star Scout: Many college students have just entered school and are still adapting to college life, or are still preparing to grind for grades in order to pursue master's or doctoral degrees, but you have spent almost all your time on robot R&D and preparing to start a business. Why did you make that choice?

Huang Yi: College is essentially a stage for trial and error at zero cost. I think it should be a "pre-society" stage, so you absolutely must not live through college as if it were high school. I did not want to evade the society I will inevitably enter, so I went straight into the lab and started a business in my junior year, hoping to get some real social feedback.

Z Star Scout: You built a drone in your freshman year. Why did you later want to build robots?

Huang Yi: Robotics is an industry that, once you step into it, you may never get out of for the rest of your life. It motivates you through positive feedback, and each generation of robots must be more challenging than the previous one. I do not want to make something without a fundamental technological breakthrough.

Huang Yi says he has a naturally freedom-loving nature, and what he longs for most is to explore freely in the endless ocean. He likes to compare himself to a "pirate captain," which means he can sail a pirate ship aimlessly in search of treasure, and along the way keep calling on brothers from all sides to come aboard and join his "treasure-hunting journey." To him, that is a very cool feeling.

Z Star Scout: I recently saw an article saying you like the title "pirate captain." Exploration is exploration, so why not explorer?

Huang Yi: Explorers have goals, such as getting over that mountain. "Pirates" have no goals and are more unrestrained. Entrepreneurship itself is like not knowing what the industry's endgame is and not knowing where the treasure is; you can only search along the way. Moreover, "pirates" have a bunch of brothers, people keep coming aboard along the way, hunting for treasure together and growing together along the way.

Z Star Scout: As a "pirate," what is the ultimate treasure you are looking for?

Huang Yi: It may be a kind of idea. For example, I believe embodied AGI will happen at RoboParty, and I believe my parents will choose to buy my products because my robots are well made. These are all treasures discovered along the way on the sea, like opening treasure chests; you never know what treasure corresponds to the next milestone.

One year of entrepreneurship, ten years' worth of words spoken, ten years' worth of hardship endured

In 2025, Huang Yi, then a junior, finished his credits and chose to graduate early. He brought several partners from Harbin Institute of Technology to Shanghai and formally founded RoboParty. Regarding how it has felt to run a startup for more than a year, Huang Yi admits that he seems to have endured ten years of hardship and spoken ten years' worth of words. The impact from the real business world has given him unimaginably rapid growth and constant iteration of self-awareness. This kind of "evolution" makes him feel as if he has lived a "compressed life."

Z Star Scout: You said that one year of entrepreneurship feels like ten years of hardship. Where does this hardship mainly come from?

Huang Yi: Mainly in terms of cognition. Three or five months ago, our understanding of models, of company management, and so on, was rapidly growing compared with today. When doing technology, the world is simple: as long as the parameters are right, it works. But running a company is completely different. Cash flow, supply chain, HR, a pile of things you could not touch in the past come rushing at you. Silicon-based information transmission is highly efficient, but person-to-person verbal communication actually has very low information density. A lot of confusion follows. After you complete one goal, sometimes you cannot clearly see where the next goal is, and this uncertainty is very tormenting.

Z Star Scout: How much did you know about the real business world in the early days of entrepreneurship?

Huang Yi: I had no concept of cost at all. It was only after getting the first round of financing that I truly began to realize that every sum of money carries responsibility. With today's understanding, I would regret not hiring finance and HR earlier. Many technical entrepreneurs neglect back-office positions, but those are the foundation of a company.

During the interview, we asked Huang Yi what the happiest moment since founding the company has been. We expected that, as a technical entrepreneur, he would describe some particularly unforgettable moment of technological breakthrough, but unexpectedly, Huang Yi believes that his happiest moments over the past year or so have all come from the colleagues in the company. Everyone is realizing their own ideas and doing things they love and that are creative together, and this feeling makes him feel very happy.

Most of RoboParty's employees are "post-00s," with interns making up nearly half, and several department heads are still master's or doctoral students in school. Besides daring to use "little geniuses," Huang Yi has also poached "veteran masters" from all sides, and even once persuaded an executive who originally held a high salary at a big company to voluntarily take a pay cut and join this startup.

Huang Yi: We have a colleague who wasn't originally in the robotics industry—he jumped ship into our robotics industry. He told me this is the job he's liked the most out of any he's done. To me, that's actually a very happy point.

Z Star Scout: What other things on the entrepreneurial journey have been very unforgettable for you, or even overturned your understanding?

Huang Yi: For example, founders should spend 30% to 50% of their time on hiring. At first we would consider hiring someone whose ability just barely sufficed for the current stage, and later we found that was wrong. For any position, the best logic is to first hire someone far above the ability required for that position—first find the No. 1 person for that position. He has deep experience in that position, and then he recruits the people after him.

Fully open source is the only way to get closer to the endgame

One of RoboParty's most eye-catching labels is full-stack open source. It not only opens the control protocols of all key modules, but also opens the module drawings, driver schematics, complete machine structure, heat dissipation, wiring harnesses, test reports, and so on—even the underlying parameters that are often regarded as core commercial moats in the industry are all open-sourced. Among the hundreds of companies in China's embodied intelligence track, this can be considered an "alternative" that lays all its cards on the table.

In Huang Yi's eyes, only open source can reject the resource waste of the entire embodied intelligence industry reinventing the wheel, and only by opening up the underlying layers and letting the vast number of developers explore forward on the existing foundation can the AGI moment of embodied intelligence as a whole arrive sooner. In this process, RoboParty's competitive core is product strength—making the product absolutely "SOTA" (the best model and method currently known in a certain task domain).

Z Star Scout: RoboParty is fully open source, so how do you make money?

Huang Yi: Open source doesn't mean not making money; it's more about whether your product is good after open-sourcing. Will everyone be willing to pay more for open source? Open source is not charity. We anchor on the scientific research market, which has very high compounding returns. Today's developers will become the builders of the entire industry in the future. By making robots open source, maybe among my users there are Bill Gates and Steve Jobs, and they will help make my robots better—to me this is also a sense of mission.

Z Star Scout: When the initial product was just open-sourced and released, how was the feedback?

Huang Yi: At that time we posted a Bilibili video less than 1 minute long, recorded with a phone, and it brought us shipments of roughly more than 100 robots per month. At that time we actually only wanted to verify product-market fit, and our generation of robots wasn't even in a fully assembled state when shipped to customers—they were shipped as parts and users had to assemble them themselves. But by a strange twist of fate, many customers now even specifically ask for parts.

Z Star Scout: Fully open source—have you never struggled or had concerns?

Huang Yi: I did agonize over whether to polish it to absolute perfection before releasing it, and later I figured it out: open source can't save a bad product. If the product itself isn't good, no matter how much you open up, no one will use it—just as if DeepSeek were an ordinary model, probably no one would care about its open source, but because it has very good performance, it constitutes the "kill line" of the large-model industry. Open source is not a pose; the premise is that the product must be done well and solidly.

Z Star Scout: Full open source requires absolute confidence—so what is your moat?

Huang Yi: The moats of the future robotics industry may lie in three categories. The first is the chip side—weighing power consumption, performance, and heat generation to strike a better balance; the second is privatized data—data for a certain scenario that only you have; the third is the brand side. Just like buying "NIO, XPeng, Li Auto" today, everyone feels there isn't much perceptual difference in the basic parameters, and people look more at brand image. In the future, embodied intelligence brands will also be a very big moat. Our real moat is not static drawings, but the ability to continuously iterate, talent density, and the entire ecosystem.

A Gundam robot secondarily developed by Huang Yi and developers

RoboParty's open source has spawned a rich ecosystem. Many developers conduct secondary development based on RoboParty's robot body, transforming it into walking Gundams and Astro Boys, which has brought Huang Yi huge surprises. He is also increasingly determined to treat the company's interior as the core circle, and then continuously broaden the ecosystem circle of the vast number of external developers through open source, so that the thriving ecosystem can better nourish all robot enthusiasts to create freely.

The ultimate romance of the "Robot Party"

The name RoboParty sounds very Gen Z, like a never-ending carnival.

Huang Yi has always believed that the communication bandwidth between robots is far greater than that between humans, and that in the future it will change organizational models and even change the whole society. In a robot bar many years from now, robots of different forms will gather together for a party, exchanging information at flying speed. Robots might even lead a robot dog on a leash, and they might even walk up to humans to introduce and sell themselves proactively. They will also land and enter everyone's home, providing for elderly parents and interconnecting with all kinds of smart devices at home... Everything humans can do, robots will be able to do in the future.

Z Star Scout: An unavoidable question—Unitree and other leading companies are already running ahead, so where is your living space?

Huang Yi: Our products are in the process of gradual mass production. From the perspective of vertical integration capability, we have done relatively heavy work—we've made key modules, small-scale manufacturing, and pilot production lines, and have taken on the heavier parts, which will bring lower costs and faster iteration speed. Our customer group has expanded the industry pie. In the past, most companies making robot hardware served algorithm engineers, but a large number of students in electronics and structural directions find it hard to get a body suitable for their own development—these people are our opportunity.

Z Star Scout: Currently, what is RoboParty's core product?

Huang Yi: We focus on one product and keep iterating. The newly launched RP1 solves many pain points in the current industry. For example, a series of problems such as pinched hands, overheating of the shoulders and ankles, and the lack of an emergency stop have all been solved by us.

At the end of the interview, we asked Huang Yi to define himself in three words. This "post-00s" CEO in the hard-tech track chose freedom, love, and romance. He yearns extremely for a free life that follows his own heart. If one day RoboParty realizes all the wishes he made for the embodied intelligence industry, he says he really will pilot a boat out to sea to find new treasures.

Huang Yi said he will never forget the moment his drone first lifted off in his freshman year—it was a pure, unadulterated joy driven by technology. "Keep looking, don't settle." The sea has no end, and treasure hunting is always on the way.