Liang Wenfeng has no personal life, Yang Zhilin has no way back

Original author: Jia Liu

In the week following the release of Kimi K3, the US tech circle was collectively lamenting why a brilliant young talent like Yang Zhilin didn't stay in America back then.

Kimi K3 surpasses all models in front-end capabilities

This topic garnered over five million views on X. David Sacks, the former White House AI lead under Trump and a close confidant, even said he has switched from Claude to Kimi for a lot of his work, "It's just more fun because it gets straight to work instead of lecturing you." Meanwhile, veteran Silicon Valley VC Vinod Khosla blamed immigration policy, stating that America is personally scaring away exceptional talent.

As speculation snowballed, Yang Zhilin's doctoral advisor, Salakhutdinov, publicly "showed off" his beloved student on social media while also clarifying the reason Yang didn't stay in the US: "If he didn't even have the courage to try starting a business, Yang Zhilin said he would regret it for the rest of his life."

On the other side of the ocean, another name from Guangdong, Liang Wenfeng, is also widely known in China's AI circle.

Seven years older than Yang Zhilin, Liang was born in Wuchuan, Zhanjiang. He's a programmer who spent fifteen years in quantitative trading, has almost never given interviews, has no social media accounts, and his colleagues' only description of his personality is "has no hobbies other than programming." After his DeepSeek R1 was released in January 2025, Nvidia's market cap evaporated by nearly $600 billion in a single day, an event Silicon Valley called the "Sputnik moment." While the whole world was looking for him, he retreated to his hometown to play soccer for a few days.

Two men from Guangdong, one born in Wuchuan and the other in Shantou, separated by the Leizhou Peninsula. Their companies are on the same track, following nearly mirror-image trajectories, yet every fork in the road stemmed from their fundamental personalities.

A Radio and a Band

Liang Wenfeng was born in 1985 in Mililing Village, Tanba Town, to parents who were both primary school teachers. With few toys at home, his most important childhood possession was a "Feiyue" brand radio, which he disassembled and reassembled countless times.

This quiet child showed his unique qualities early on. His middle school homeroom teacher remembers he wasn't a bookworm or necessarily more diligent than others, but he had taught himself high school math by junior high and started browsing college textbooks, "as if he could master every subject without spending much time."

In the 2002 college entrance exam, he scored 806 points (out of 900), becoming the top scorer in Zhanjiang City. A photo from the award ceremony is still searchable: wearing a maroon short-sleeve shirt with a large red flower pinned to his chest, his expression stiff, seemingly pushed onto the stage by a teacher. That autumn, he entered Zhejiang University to study Electronic Information Engineering. But for the next twenty years, this man would hardly ever be in a situation worth photographing again.

Seven years later in Shantou, another boy from Guangdong grew up. Yang Zhilin, born in 1992, ranked first in his class for all four years in Tsinghua's Computer Science department and published over twenty papers. In Tsinghua, this wasn't the most extraordinary part. What was unusual was that, amidst these achievements, he formed a rock band named Splay (after a data structure) and played the drums himself.

He later explained, "At the time, I felt I had a lot to express, including pressure from reality and the absurdity of the broader environment." They wrote a song about the daydream of getting rich overnight through a successful startup, "half out of empathy, half as a reminder to myself not to become too utilitarian."

Many years later, this band would re-enter the story in a way no one could have predicted: bandmate Zhou Xinyu became a co-founder of Moonshot AI.

And in the office of Moonshot AI, a white Yamaha electric piano sits at the entrance, holding Pink Floyd's 1973 album "The Dark Side of the Moon" (Yue zhi An Mian). The company name comes from this very album.

The Dark Side of the Moon album

One dismantles radios needing no audience; one plays drums because he has something that *must* be expressed. Nearly every choice these two men made over the following two decades stems from this core difference.

A Rented Room in Chengdu and the Halls of CMU

After graduating from Zhejiang University, Liang Wenfeng didn't go to a big tech company for a technical credential. He went to Chengdu, hid in a cheap rented room, and tried various algorithms, aiming to equip traditional industries with AI. It all failed. Most founding teams of China's top quantitative funds had gilded resumes from overseas hedge funds; Liang figured it out on his own in that rented room.

In 2015, he co-founded High-Flyer with a Zhejiang University classmate. A series of moves followed that almost no one understood at the time: in 2019, he spent nearly 200 million RMB to build an in-house cluster with 1100 GPUs; in 2021, he added another billion, stockpiling about 10,000 A100 GPUs.

You didn't need that many cards for quantitative trading. Liang himself admitted that for trading alone, a few cards were sufficient. A former acquaintance recalled that when seeing Liang stockpile cards to train models, he just thought it was a bad-haired tech geek burning money.

But what he was doing was the opposite of what everyone assumed. He wasn't using AI to reduce costs and increase efficiency in finance; he was using finance to fuel AI research. Most Chinese AI companies follow the sequence of getting funding first, then finding a product, then finding cash flow. He completely reversed it: build a cash-flow machine first, then use it to buy the freedom to do research.

Yang Zhilin walked a different path, one paved with flowers and landmines.

After graduating from Tsinghua, he went to Carnegie Mellon University (CMU) for his PhD, during which he did research at Google's and Meta's AI labs. In 2017, he bet all his energy on language models, later calling it "the only important problem." During his PhD, he published two papers: one taught AI to remember longer contexts, and the other defeated Google's own strongest model in 20 tests. Together, they have garnered nearly 20,000 citations.

Years later, when Kimi became popular for its ability to handle long text inputs, many thought it was a last-minute differentiating feature found in 2023. In reality, it was just a different form of the direction he had identified during his PhD.

In 2016, while still a PhD student, he co-founded Recurrent AI, focusing on sales call analysis, with Sequoia Capital China and GSR Ventures as shareholders. This early venture gave him a firsthand look at the messiness of technology implementation but also planted a landmine under his feet, which would detonate eight years later.

Upon graduating with his PhD in 2019, his advisor connected him with an Apple executive who reported directly to Tim Cook, asking if Yang would like to join Apple, even potentially at Apple China. Yang turned down the email from Apple and offers from Silicon Valley, deciding to return to China.

At that time, Liang Wenfeng was stockpiling GPUs in Hangzhou, and Yang Zhilin was waiting for the wind to pick up in Beijing. Neither knew of the other's existence, nor that these two names would come to represent today's Chinese AI circle.

A One-Month Window and a Catfish

On November 30, 2022, ChatGPT was launched, causing collective insomnia in Silicon Valley's tech circles. Yang Zhilin recalled that many friends around him were anxious, experiencing FOMO, and unable to sleep. Many turned around and started companies.

"We started focusing on the first round of fundraising in February 2023; if we had delayed until April, the opportunity would have been mostly gone. But doing it in December 2022 or January 2023 wouldn't have worked either, due to the pandemic; people hadn't reacted yet." Yang Zhilin seized that one-month window. He didn't rest for a single day.

In March 2023, Moonshot AI was founded, with co-founders being Tsinghua alumni Zhou Xinyu and Wu Yuxin. It reportedly secured $60 million in initial funding and gathered about 40 AI researchers in three months. Then came one of the steepest fundraising curves in the history of Chinese large models: Sequoia Capital China and ZhenFund invested, Alibaba led a $1 billion round, followed by Tencent, Meituan, and Xiaohongshu, pushing the valuation to $3.3 billion. Kimi, with its 200,000-character long-text capability, became the first large-model product used frequently by many Chinese people.

During this time, Yang Zhilin lived in a dual state. Externally, he told the grandest narratives, estimating the probability of scaling laws failing as close to zero, comparing starting the company to driving towards endless snowy mountains, and calling the first year building a rocket prototype and getting a handle on the fuel formula. Internally, he had to focus on the most trivial algorithmic details: when computing power was tight, the cost for one machine fluctuated daily – 260 one day, 340 the next, dropping back a few days later. Whether to buy or rent, which channel to use – he tracked and adjusted daily. The certainty of a scientist and the shrewdness of a small business owner were both embodied in this 31-year-old.

But everything capital provided came with a price tag attached. To sustain the growth curve, Kimi spent 220 million RMB on user acquisition in October 2024, and another 200 million in November, burning through more than the entire Q3 budget in two months. The drummer who wrote a song satirizing overnight riches became the founder spending the most aggressively on user acquisition in the entire industry. It wasn't that he had changed, but the $3.3 billion valuation was making the decisions for him.

Liang Wenfeng entered the scene in Hangzhou in a way that seemed almost deliberately contrary.

In 2023, DeepSeek was spun off from High-Flyer, refusing any external investment. The team had fewer than 140 people, almost no returnees from overseas, consisting mostly of recent graduates from Chinese universities and young people a few years out of school. No KPIs, no hierarchy. Those with ideas could directly request computing resources and personnel.

A former employee told *The Washington Post* that Liang Wenfeng would delve into the details of training strategies, reading papers and writing code with the researchers, "completely unlike a boss, more like a geek." He explained why he hired fresh graduates instead of poaching industry stars: "Experienced people will tell you what to do without thinking, but inexperienced people will explore and experiment repeatedly."

In May 2024, DeepSeek-V2 slashed its API price to 1 yuan per million tokens, forcing ByteDance, Alibaba, Baidu, and Tencent to follow suit. The whole industry thought it was a long-planned business war tactic. His response was: "We didn't intend to be a catfish; we just accidentally became one."

Pricing was just a slight margin above cost, "not subsidizing, not making exorbitant profits." Internet people talk about market share, entry points, and network effects during price wars; he talks about cost accounting. But it's precisely this emotionless price cut that is most deadly, dragging the narrative of large model APIs from high-margin storytelling directly into the pricing logic of infrastructure.

At some point, the outside world began comparing Yang Zhilin and Liang Wenfeng's two completely different business models.

Yang Zhilin in Suspension

The landmine Yang Zhilin had buried eight years ago exploded in November 2024.

Five old shareholders from his previous venture, Recurrent AI, filed for arbitration in Hong Kong, accusing him of starting fundraising for *Moonshot AI* without obtaining waivers from all shareholders of the previous company.

On December 5th, Zhu Xiaohu (a prominent Chinese VC) launched a scathing attack on his WeChat Moments, targeting Zhang Yutong: the former GSR Ventures partner allegedly received 9 million shares (an initial 14% stake) in Moonshot AI for free, surpassing the 9.5% stake allocated to the "parent company" Recurrent AI. Zhu's proposed solution was almost humiliating: apologize, return the shares, or have the company cut ties with Zhang Yutong.

At 9:40 PM on December 6th, Yang Zhilin posted a 1,300-character statement. He didn't cut ties; instead, he drew a firm line: Zhang Yutong was a co-founder, the shares were compensation for years of future work, and the procedures for leaving Recurrent AI had been signed off by every director. People close to the company relayed the internal sentiment: she is already one with Moonshot AI; can't be cut off.

Zhu Xiaohu publicly expressed complete incomprehension. In a purely commercial coordinate system, it was indeed a dead end; cutting ties was the only rational option. But Yang Zhilin's decision-making coordinate system included other considerations. Salakhutdinov's later clarification served as a footnote: this is the kind of person who "would regret it for a lifetime if they didn't try." Once he decides on something, he doesn't look back, even if the cost is already on the table.

The real hammer fell forty-some days later.

On January 20, 2025, R1 was released. Free, open-source, with reasoning capabilities rivaling OpenAI's o1. A team of just over a hundred people in Hangzhou turned the global capital markets upside down within a week. Matt Sheehan, a researcher at the Carnegie Endowment, made a very interesting remark: "DeepSeek wasn't the company China pre-selected; its explosive success even surprised China."

Meanwhile, Liang Wenfeng was celebrating the New Year back in his hometown, Wuchuan. On the afternoon of January 27th, he played a game of soccer in the village with his middle school classmates. The village entrance was crowded with tourists checking in, while the protagonist himself was on the soccer field.

For Yang Zhilin, this was a double blow. The arbitration was unresolved, and R1 directly sentenced his strategy of the past year to death: users acquired by spending heavily were insignificant against a free and stronger competitor. Public opinion turned. The title of one article read: "Yang Zhilin, the Suspension of a 90s Idealist." Suspension meant his feet weren't on the ground. When he wrote songs, he worried about becoming utilitarian; now the world was saying he was both utilitarian and a failure.

At the beginning of 2025, Moonshot AI still had money, but very little narrative power.

The Comeback Against Gravity

The next year was crucial for Yang Zhilin.

Yang Zhilin essentially negated his own actions of the previous year. He stopped user acquisition spending, cut redundant business lines, streamlined operations to focus on the foundational model, and shifted to open source. In a conversation at GeekPark, he said organizational inertia pushes you to do more and more things. "We have to fight this gravity."

It's easy to say, but doing it means admitting the previous path was wrong, laying off the people you hired, and bowing your head to a competitor that nearly destroyed you. Most 33-year-old founders can't get over this hurdle. Yang Zhilin did it with unusual decisiveness. Perhaps open-source and long-termism were always his default settings; closed-source and aggressive user acquisition were the clothes capital dressed him in, and he was just taking them off now.

In July 2025, the trillion-parameter K2 was open-sourced. In November, K2 Thinking surpassed GPT-5 on several of the hardest Agent benchmarks. Hugging Face co-founder Thomas Wolf asked on Twitter: "Is this another DeepSeek moment?"

In the early hours after the release, Yang Zhilin, along with Zhou Xinyu and Wu Yuxin, hosted an AMA on Reddit, answering 21 questions. They clarified that the $4.6 million training cost figure was not official, acknowledged having fewer GPUs than US counterparts, "but we squeeze the absolute maximum performance out of every single card." Someone asked how they viewed OpenAI burning money. Zhou Xinyu replied casually: "We don't know either, only Sam knows... we have our own rhythm."

"Our own rhythm." In 2024, Moonshot AI couldn't have said those five words. Its rhythm then was the investors' rhythm, the ROI of user acquisition rhythm. The one who gave these five words back to Yang Zhilin was precisely Liang Wenfeng. R1 proved that open-source plus algorithmic efficiency could work in China, effectively acting as a roadshow to *his own board of directors*. The person nearly killed by DeepSeek ended up having to thank it for