A VC's ten-year growth journey: achieving a unicorn in 3 years and reaching a trillion-dollar market cap with an IPO in 10 years

Author | Xue Haohao

Editor | Barry

Image Source | Yunqi Capital

In early January 2026, MiniMax rang the bell at the Hong Kong Stock Exchange, surpassing Zhizhuo and reaching a market value of over HKD 100 billion. On March 10, driven by the “Lobster Farming” (i.e., open-source AI agent framework OpenClaw) craze, MiniMax’s market value reached approximately HKD 382.6 billion, marking its first time surpassing Baidu.

Four years ago, when the company was just founded, Chen Yu from Yunqi Capital participated in MiniMax’s first angel round investment, with a post-investment valuation of only 1.2 billion yuan.

Chen Yu (left) and Yan Junjie (center) at the MiniMax listing scene

From distributed databases in 2016 to foundational large models in 2021, Chen Yu, with 12 years in the industry, has made a series of precise investment decisions, backing unicorns like MiniMax, Yuanrong Qihang, New Stone Age Unmanned Vehicles, PingCAP, and others. These projects all entered at the angel or Series A stage early on. Currently, he has become the third managing partner at Yunqi Capital, after its two founders.

“I see myself more as a hunter. Whenever I enter a new track, I conduct in-depth research, and I like investing in smart, honest, ambitious young people,” Chen Yu says.

Chen Yu believes that becoming a qualified VC investor requires crossing three time thresholds—3 years, 7 years, and 10 years.

“A competent investor, after 3 years in the industry, should have backed a unicorn; by 7 years, that unicorn should still be alive; by 10 years, the unicorn should go public and bring substantial returns to the fund.”

This “3, 7, 10” rule is his “upgrade guide” for new venture capital managers.

Year 3

Make your first unicorn

Chen Yu emphasizes to CY Zone that one should invest when most investors don’t understand or see through the opportunity.

He didn’t enter the investment industry right after college but worked as CTO at Limei Technology, an intelligent marketing platform startup. However, even before becoming CTO, he had plans to become a tech investor.

His purpose in joining Limei Technology as CTO was: “To do good investments, you need to see clearly what the person on the other side of the negotiation table has experienced and is doing. Knowing yourself and your opponent allows for better investment decisions.”

Data shows Chen Yu graduated with a bachelor’s degree in computer science from Sun Yat-sen University, earned a master’s in computer science from Johns Hopkins University, worked for years in technical development at Google, and then pursued an MBA at the University of Chicago.

Notably, during his MBA studies, Chen Yu interned at IDG Capital, where the team leader was Mao Chengyu, the founder of Yunqi Capital.

Founder of Yunqi Capital Mao Chengyu

In 2014, Mao Chengyu left IDG Capital and co-founded Yunqi Capital with Huang Yubin. By then, Chen Yu had already gained three years of experience as CTO at Limei Technology (which had raised $10 million from IDG in 2011, another $20 million from KPCB in 2012, and in 2015, secured RMB 20 million from Yunqi and GaoRong Capital, and went public on the New Third Board in 2016), and was preparing to enter the investment field. The newly established Yunqi Capital opened the door for him.

With a background in hard tech and professional business training, Chen Yu made a significant investment just two years after joining Yunqi: PingCAP.

In early 2016, a company called PingCAP faced cold reception during Series A funding. “Basically, no investors understood what they were doing at the time.” The company develops distributed databases.

At that time, most large-scale financing in China focused on innovative business models in internet applications. The top five funded companies were Ant Financial, Ele.me, Didi, Meituan-Dianping, and JD Finance, each raising billions of dollars. In contrast, distributed databases—a hard tech track—were considered a “niche.”

Moreover, PingCAP’s founder Liu Qi was from Yangtze University, not Tsinghua or Peking University. In China’s VC industry, that’s usually a disadvantage rather than an advantage.

Chen Yu was different. He still intended to connect with Liu Qi and understand his “technical fundamentals.”

Through an introduction from a peer at a US dollar fund, Chen Yu met Liu Qi in person. Starting from his experience in Google’s technical R&D, Chen Yu found: “PingCAP’s product ambitions are very high—they aim to compete with Google’s core storage systems—and I think their technical taste is very good.”

That initial conversation was just a preliminary judgment. What truly convinced Chen Yu to invest was his examination of PingCAP’s underlying code. During due diligence, he specifically reviewed the code—not only the code itself but also the architecture design and thought process behind it.

This reaffirmed his view: PingCAP has a deep understanding of the essence of distributed databases. “It’s difficult, but they have a chance.”

“Many teams from Tsinghua or Peking University don’t necessarily write better code than Liu Qi,” Chen Yu says. He’s seen many products from top universities that are mediocre, and others from ordinary universities that are genius-level. “So I never judge based on academic background.”

In 2016, Chen Yu led Yunqi Capital’s Series A investment in PingCAP. The company’s subsequent development validated his judgment. In 2018, PingCAP’s $50 million Series C set a record for funding in China’s distributed database sector. Today, PingCAP is a unicorn valued at over $1 billion.

Investing in PingCAP was Chen Yu’s first unicorn deal, in his second year in VC, aligning with his “3, 7, 10” rule’s first stage.

“A competent VC, after 3 years, should have backed a unicorn; after 7 years, that unicorn should still be alive; after 10 years, it should go public and generate substantial returns.”

Year 7

Is your unicorn still alive?

After investing in PingCAP, Chen Yu began to refine his investment logic and subsequently backed two more unicorns in autonomous driving—Yuanrong Qihang and New Stone Age Unmanned Vehicles.

He repeatedly emphasizes the ideal founder profile: smart, honest, technically tasteful, resilient, and ambitious.

“Intelligence increases the likelihood of doing things well; honesty helps objectively analyze setbacks without making short-sighted decisions; technical taste helps founders anticipate technological trends early; resilience allows entrepreneurs to persist through difficulties; ambition aims for big results that can bring substantial returns to investors,” Chen Yu explains.

Compared to other early-stage investors who emphasize “people,” Chen Yu places greater importance on the “thing” founders are doing—namely, the product and technology.

Around 2016, autonomous driving was a hot track. Tesla, Waymo, Baidu, and others were betting heavily. Leading startups like Xiaopeng, Wenzhou Zhixing, and RoadStar had already begun to emerge. Yunqi Capital invested in RoadStar at the angel stage. The three core founders—Tong Xianqiao, Heng Liang, and Zhou Guang—had previously worked on localization, perception, and calibration in Baidu’s autonomous vehicle project.

However, within this hot track, RoadStar faced a critical problem: internal conflicts among the founders grew intense and erupted in 2019.

Tong Xianqiao and Heng Liang issued a joint statement to oust Zhou Guang, accusing him of secretly opening code repositories and accepting kickbacks. Later, the investors collectively denied the statement’s validity and supported Zhou Guang. The founder team tore itself apart, and RoadStar eventually went bankrupt.

Each of the three founders chose to start anew, raising funds externally. Amid this chaos, Chen Yu made a bold decision: to invest in Zhou Guang’s new venture—Yuanrong Qihang.

This decision was difficult because all three founders had strong technical backgrounds and were high-profile entrepreneurs. It was hard to judge who was better.

“To decide which founder to invest in, we interviewed 30-40 employees from RoadStar, accounting for about a third of the company’s staff. We judged Zhou Guang to be the most technically skilled and to have made the greatest contribution,” Chen Yu says. “So we chose to invest in Zhou Guang’s Yuanrong Qihang, especially since autonomous driving is a highly technical track.”

Today, Yuanrong Qihang has become the third company after Huawei and Momenta to achieve high-level autonomous driving with NOA (Navigate on Autopilot) in urban environments, ascending into the top tier of driverless solutions. It has completed seven funding rounds and is valued at unicorn level.

In the same year he invested in Yuanrong Qihang, Chen Yu planned to invest in New Stone Age’s Series A. Yuanrong Qihang focuses on high-speed passenger autonomous driving, while New Stone Age develops low-speed logistics vehicles for urban streets, industrial parks, and logistics centers—most notably, last-mile delivery in urban logistics.

However, just as he was about to invest, he discovered that Zhu Lei, a core member of New Stone Age, had left to start Bai Xiniu, becoming a competitor.

The departure of a core member and emergence of a rival posed a potential threat to New Stone Age’s development. Should he continue investing or give up?

Through due diligence and professional judgment, Chen Yu found that Zhu Lei’s departure had limited impact on New Stone Age’s growth. He decided to proceed with the Series A investment and increased his stake in the A+ round.

New Stone Age’s subsequent development was also fraught with difficulties. “Founder Yu Enyuan experienced two major core team departures, funding shortages, market skepticism, and more—enough to fill a business novel,” Chen Yu notes.

At the toughest moment, Yu Enyuan even mortgaged his house to keep the company afloat.

From its founding in 2018 to today, Chen Yu waited seven years to see New Stone Age’s rise.

What happened during these seven years? The company was slowly accumulating strength.

From 2018 to 2024, New Stone Age’s deployment of unmanned vehicles increased gradually. But in 2025, the situation changed dramatically. With technological maturity and open road rights, “the company deployed 10,000 unmanned vehicles in 2025, compared to only 2,000 before.”

Chen Yu says: “Yu Enyuan is a resilient entrepreneur who thrives in adversity—that’s a trait I highly value.”

“Many investors look at a startup and ask how big the future market is, how many users it will have. But that’s not the key. The real question is: when will the technology reach a critical point? Once it does, how big can the market become?” Chen Yu explains.

From PingCAP to Yuanrong Qihang and New Stone Age, Chen Yu has backed multiple unicorns within just five or six years, and these companies remain leaders in their respective tracks.

This again confirms his own theory: a qualified VC, after 7 years, will still have invested in companies that are alive and thriving.

“Many so-called unicorns are fake,” Chen Yu adds. “Only truly valuable unicorns can survive long-term.”

Year 10

Reaping a Super IPO

As an angel investor in MiniMax, Chen Yu entered the Hong Kong stock market in early 2026, successfully becoming the second AI large model company after Zhizhuo to go public, with a latest market cap exceeding HKD 300 billion.

“Projects like MiniMax in their early stages are usually accessible only to a few specific funds or those with close relationships to the founders. Most investors don’t have access to the project or the founders, so they can’t judge the quality. Many don’t even know such projects exist,” Chen Yu explains. “Investors who get early information can make better decisions through professional judgment.”

MiniMax’s Series A was completed efficiently at the end of 2021, with founder Yan Junjie meeting only a few closely related investors, raising $31 million. Investors included Yunqi Capital, IDG Capital, MiHoYo, and Hillhouse Capital.

“At that time, many entrepreneurs were still considering small models to solve problems, but Yan Junjie decided to use a more general foundational model. I agree with his vision and persistence for future AI technology,” Chen Yu says.

By late 2022, investing in AI large models was not yet industry consensus. But by mid-2023, all major players in AI large models had emerged, quickly becoming industry leaders. These top AI model companies had mostly completed early funding rounds, with valuations soaring.

Chasing the trend and investing only when the market is hot means missing the best timing. “That’s why most investors miss opportunities—they wait for consensus to form, but by then, the best investment window has already closed.”

When Chen Yu invested in MiniMax at the end of 2021, he was promoted internally at Yunqi Capital to partner. In early 2026, as MiniMax went public, Chen Yu also achieved a significant promotion, becoming the third managing partner at Yunqi Capital after founders Mao Chengyu and Huang Yubin.

All this confirms the essence of early-stage investing: before consensus forms, investors must rely on independent judgment. And the source of independent judgment is a deep understanding of technology.

The unicorn founders Chen Yu invested in are not all from illustrious backgrounds: Liu Qi from Yangtze University, Yu Enyuan from Chongqing University, Zhou Guang who graduated from Tsinghua but was not outstanding academically, Yan Junjie from Southeast University. Yet, they all possess the “technical taste” that Chen Yu values.

What is “technical taste”?

The ability to analyze and judge a founder’s technical and product quality through their background, and whether they can grasp the mainstream technological development path in the future.

“When many investors rely on prestigious backgrounds like Tsinghua, Peking University, Stanford, and shiny career resumes for investment decisions, founders with more modest backgrounds but high potential find it hard to get enough funding. For someone like me who understands technology, that’s a good thing.”

Take MiniMax as an example. When Chen Yu first met Yan Junjie, they talked all night about technical issues.

“At that time, many AI entrepreneurs would train a small AI model to solve a specific problem. Yan Junjie chose to train a foundational large model to handle most problems. That’s a more efficient approach, more aligned with how the human brain (closer to AGI) works. He has the technical taste I recognize,” Chen Yu says.

This “technical taste” also guided all subsequent technical route choices for MiniMax.

Later, Yan Junjie led the team to become one of the earliest domestic AI large model companies to adopt MOE (Mixture of Experts) architecture (DeepSeek later also adopted a similar architecture), and in recent years, they have begun to focus on linear attention mechanisms. The ability to handle infinite-length texts via linear attention is a key technical dimension for MiniMax, crucial for its social applications, emotional companionship, and Agent use cases.

Therefore, understanding technology is extremely important for investors in the current hard tech era. “Now, to recruit investment managers, they must understand technology. Candidates without a STEM background can’t even pass the first resume screening,” Chen Yu states.

“He who understands technology can communicate on the same wavelength with founders, especially technical founders. If you don’t understand technology, founders won’t trust you or give you investment opportunities. Many key pieces of information won’t be shared, making it hard for you to make complete investment judgments,” he adds.

Besides valuing founders’ technical taste, Chen Yu also emphasizes their business intuition.

Initially, Yan Junjie chose to focus on international markets, multi-modal AI large models, and To C markets. Because only by doing so could the future market potential be high enough.

“When valuing MiniMax, I don’t set an absolute valuation—say, I won’t invest if it exceeds 500 million RMB. Instead, I focus on relative returns. If I believe it will grow to a valuation of hundreds of billions, then even a $200 million investment isn’t expensive,” Chen Yu explains.

This “relative return” mindset allows him to enter at seemingly “high” valuations early on, ultimately earning dozens of times returns.

Beyond potential market size, he also pays attention to the company’s commercialization progress. “Egg-laying along the way is important: continuously doing business, developing products or services, gathering market data, validating technology, and adjusting direction.”

Take Yuanrong Qihang as an example. Chen Yu believes that shifting from L4 to L2+ in 2022 was not abandoning technological ideals but recognizing that L4 deployment costs were too high. The strategic adjustment accelerated their commercial deployment of autonomous driving technology. “Commercialization isn’t something to consider later; it’s something to think about from day one.”

A good investor doesn’t just occasionally hit good projects but can consistently invest in good ones.

Besides MiniMax, projects like Yuanrong Qihang and New Stone Age Unmanned Vehicles are also preparing for secondary market listings. Chen Yu is entering his IPO harvest phase.

Looking ahead, he is paying attention to two main tracks: “In AI hardware, China has advantages over the US—not only humanoid robots but also autonomous driving, lawnmowers, pool cleaners, and other sectors; in AI Agents, due to lower willingness to pay for software in China, we will focus more on outbound AI Agent companies.”

From entering the industry in 2014, investing in PingCAP in 2016, and then in MiniMax in 2021, Chen Yu has spent 12 years validating his “Investor Progress and Advancement Guide.”

His “3, 7, 10” rule is not just a timeline but a comprehensive VC growth logic: Year 3, make your first move before the crowd understands; Year 7, verify the unicorn’s strength and your independent judgment; Year 10, harvest a super IPO and generate large returns, completing the leap.

This article is an original from CY Zone. Unauthorized reproduction is prohibited. CY Zone reserves the right to pursue legal action. For reprints or inquiries, please contact: editor@cyzone.cn.