Liang Wenfeng Has No Life, Yang Zhilin Has No Way Out

Written by: Jia Liu, Zhang Sheng BeatZ

In the week following the release of Kimi K3, the American tech community collectively lamented why such an outstanding young person like Yang Zhilin did not stay in the United States.

Kimi K3 surpasses all models in front-end capabilities

This topic garnered over five million views on X. David Sacks, the former head of AI at the White House under Trump and a close confidant of Trump, stated that he had switched from Claude to Kimi for handling a large amount of work, saying, "It's just much more fun because it gets the job done instead of lecturing you." Meanwhile, veteran Silicon Valley venture capitalist Vinod Khosla attributed the issue to immigration policies, claiming that the U.S. is scaring away exceptional talent.

Speculation grew, and Yang Zhilin's PhD advisor, Salakhutdinov, publicly showcased his student on social media while clarifying the reasons why Yang did not stay in the U.S.: "If you don't even have the courage to try entrepreneurship, Yang Zhilin said he would regret it for a lifetime."

On the other side of the ocean, another name from Guangdong is well-known in the Chinese AI circle: Liang Wenfeng.

Liang Wenfeng is seven years older than Yang Zhilin, born in Wuchuan, Zhanjiang. He is a programmer who has been engaged in quantitative trading for fifteen years, rarely giving interviews and having no social media accounts. His colleagues summarize his personality as "having no hobbies other than programming." After the release of his DeepSeek R1 in January 2025, Nvidia lost nearly $600 billion in market value in one day, marking what Silicon Valley called the Sputnik moment. While the whole world was looking for him, he hid back at home to play soccer for a few days.

Two Guangdong men, one born in Wuchuan and the other in Shantou, separated by the Leizhou Peninsula. Their companies are on the same track, following almost completely mirrored trajectories, with each divergence stemming from their inherent personalities.

A Radio and a Band

Liang Wenfeng was born in 1985 in Mililing Village, Qinba Town, with both parents being teachers at the local primary school. With few toys at home, his most important childhood possession was a Feiyue radio, which he took apart and reassembled countless times.

This quiet child showed differences early on. His middle school teacher remembered him not as a bookworm, nor particularly hardworking, but he self-studied high school mathematics by middle school and began flipping through university textbooks, "as if he could learn every subject well without spending much time."

In the 2002 college entrance examination, he scored 806 points, becoming the top student in Zhanjiang. Photos from the award ceremony can still be found today: wearing a maroon short-sleeve shirt, a big red flower pinned to his chest, with a stiff expression, clearly pushed onto the stage by teachers. That autumn, he entered Zhejiang University to study electronic information engineering. However, in the following twenty years, he never appeared in any situation worth photographing again.

Seven years later in Shantou, another Guangdong boy grew up. Yang Zhilin, born in 1992, ranked first in his class in the Tsinghua Computer Science department, with over twenty papers published. These achievements are not the most remarkable at Tsinghua; what stands out is that he formed a rock band called Splay, named after a data structure, and played the drums.

He later explained, "At that time, I felt there were many things I wanted to express, including the pressure from reality and the absurdity of the environment." They wrote a song about the daydream of becoming rich overnight through entrepreneurship, "half out of empathy, half as a reminder to myself not to become too utilitarian."

Many years later, this band would return to the story in an unexpected way: bandmate Zhou Xinyu became a co-founder of Moonlight Dark.

The office of Moonlight Dark has a white Yamaha digital piano at the entrance, with Pink Floyd's 1973 album "The Dark Side of the Moon" resting on it. The company's name comes from this album.

The Dark Side of the Moon album

One person disassembling a radio without needing anyone to see, another playing drums because there is something that must be expressed. Every choice these two made over the next twenty years almost grew out of this.

A Rental in Chengdu and Hallways at CMU

After graduating from Zhejiang University, Liang Wenfeng did not go to a big company to get a technical business card. Instead, he went to Chengdu, hiding in a cheap rental to try various algorithms, aiming to integrate AI into traditional industries, but it all failed. Most founding teams of leading quantitative funds in China have gilded resumes from overseas hedge funds, while Liang Wenfeng learned everything by himself in a rental.

In 2015, he co-founded Huanfang with his classmates from Zhejiang University. What followed was a series of actions that almost no one understood at the time: in 2019, he invested nearly 200 million to build a self-owned cluster with 1,100 GPUs; in 2021, he added another billion, hoarding about ten thousand A100s.

Using so many cards for quantitative trading is unnecessary, and Liang Wenfeng himself admitted that for trading, a small number of cards would suffice. Those who had dealings with him in earlier years recalled that when they saw him hoarding cards to train models, they thought he was just a tech geek with a bad haircut burning money.

But what he was doing was exactly the opposite of what everyone understood; he was not using AI to reduce costs and increase efficiency in finance, but rather using finance to fund AI research. Most Chinese AI companies follow the sequence of first securing financing, then finding products, and finally seeking cash flow, but he reversed the order: first creating a cash flow machine, then using it to buy the freedom to conduct research.

Yang Zhilin took a different path, one paved with flowers and mines.

After graduating from Tsinghua, he went to CMU (Carnegie Mellon University) for his PhD, during which he conducted research at AI labs at Google and Meta. In 2017, he focused all his energy on language models, later stating that this was "the only important issue." During his PhD, he published two papers, one teaching AI to remember longer contexts and another that outperformed Google's strongest model in 20 tests, with nearly twenty thousand citations combined.

Years later, Kimi gained popularity for its ability to input long texts, and many believed this was a differentiating selling point found in 2023, but in fact, it was a direction he had identified during his PhD, just taking on a different form.

In 2016, while still a PhD student, he co-founded Loop AI, focusing on sales call analysis, with Sequoia and GGV as investors. This startup gave him an early glimpse of the roughness of technology landing, but it also buried a mine beneath his feet, which would only be detonated eight years later.

In 2019, after graduating with his PhD, his advisor connected him with an Apple executive who reported to Cook, inquiring if Yang Zhilin would be interested in joining Apple, even in the China region, but Yang Zhilin turned down Apple's email and offers from Silicon Valley, deciding to return to China.

At that time, Liang Wenfeng was hoarding cards in Hangzhou, while Yang Zhilin was waiting for opportunities in Beijing. Neither of them knew of the other's existence, nor did they know that these two names would represent today's Chinese AI circle.

A Month's Window and a Catfish

On November 30, 2022, ChatGPT launched, causing sleepless nights in Silicon Valley's tech circle. Yang Zhilin recalled that many friends around him were anxious, experiencing FOMO, and unable to sleep, with many turning to entrepreneurship.

"Starting in February 2023, we focused on the first round of financing, but it was delayed until April, and there was basically no opportunity left. However, there was no chance in December 2022 or January either, as there was a pandemic, and everyone didn't react in time." Yang Zhilin seized this one-month window, not taking a single day off.

In March 2023, Moonlight Dark was established, co-founded by Tsinghua alumni Zhou Xinyu and Wu Yuxin, reportedly starting with $60 million in financing, gathering about 40 AI researchers in three months. Then came the steepest curve in the history of financing for large models in China: Sequoia and ZhenFund entered the scene, with Alibaba leading a $1 billion investment, followed by Tencent, Meituan, and Xiaohongshu, pushing the total to $3.3 billion. Kimi became the first large model product frequently used by many Chinese people, thanks to its 200,000-word long text capability.

During that time, Yang Zhilin lived in a dual state. Externally, he spoke of grand narratives, estimating the probability of scaling law failing to be close to zero, comparing entrepreneurship to driving towards a vast snow mountain, and referring to the first year as having built a rocket prototype and touched upon a fuel formula. Internally, he had to focus on the most trivial algorithms: when computing power was tight, one machine cost 260 today, 340 tomorrow, and then dropped back down in a few days, whether to buy or rent, which channel to go through, adjusting daily, modifying daily. A blend of a scientist's certainty and a small business owner's shrewdness was embodied in this 31-year-old young man.

However, everything provided by capital had already been priced. To sustain the growth curve, Kimi planned to spend $220 million in October 2024 and another $200 million in November, burning through more than the entire third quarter in just two months. The drummer who wrote songs mocking overnight wealth became the most aggressive founder in the industry for buying traffic. It wasn't that he changed; it was the $3.3 billion valuation making decisions for him.

Liang Wenfeng entered the scene in Hangzhou in a nearly contrarian manner.

In 2023, DeepSeek separated from Huanfang, avoiding any external investment. The team has fewer than 140 people, almost no returnees, all fresh graduates and young people from domestic universities. There are no KPIs, no hierarchy; if you have an idea, you can directly adjust cards and personnel.

Former employees recalled to The Washington Post that Liang Wenfeng would delve into the details of training strategies, reviewing papers and writing code with researchers, "not at all like a boss, but more like a geek." He explained why he hired fresh graduates instead of competing for industry veterans: "Experienced people will unthinkingly tell you what you should do, but inexperienced people will explore repeatedly."

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

Pricing is merely a slight profit above costs, "not losing money, nor making exorbitant profits." While internet entrepreneurs discuss market share, entry points, and network effects in price wars, he focuses on cost accounting. Ironically, this unemotional price cut is the most damaging, dragging the narrative of high-margin large model APIs directly into the pricing logic of infrastructure.

Yang Zhilin and Liang Wenfeng, two completely different business models, have been compared by the outside world for an unknown period.

The Suspended Yang Zhilin

The bomb Yang Zhilin planted eight years ago exploded in November 2024.

In his last entrepreneurial project, the five old shareholders of Loop Intelligence initiated arbitration in Hong Kong, accusing him of starting financing for a new company before obtaining waivers from all shareholders.

On December 5, Zhu Xiaohu fired shots on social media, focusing fire on Zhang Yutong: the former partner at Jina Capital obtained an initial 14% of 9 million shares for free, exceeding the 9.5% allocated to Loop Intelligence as the "parent company." Zhu Xiaohu's solution was almost humiliating: apologize, withdraw shares, or the company would cut ties with Zhang Yutong.

At 9:40 PM on December 6, Yang Zhilin released a 1,300-word statement. He did not cut ties; instead, he made it clear: Zhang Yutong is a co-founder, and her shares are compensation for years of future work, with procedures for leaving Loop signed by every director. Those close to the company relayed the internal attitude: she and the Dark Side of the Moon have become one, and cannot be separated.

Zhu Xiaohu publicly stated he completely did not understand. In a purely commercial coordinate system, it is indeed unsolvable; cutting ties is the only rational option. But Yang Zhilin's decision-making coordinate system has other considerations. Salakhutdinov's later clarification provided a footnote: this is the kind of person who believes "not trying will lead to lifelong regret," and once committed, does not turn back, even if the cost is already on the table.

The real heavy blow fell more than forty days later.

On January 20, 2025, R1 was released. Free, open-source, with reasoning capabilities rivaling OpenAI's o1. A team of over a hundred in Hangzhou turned the global capital market upside down within a week. Carnegie researcher Matt Sheehan said something very interesting: DeepSeek was not the company chosen by China in advance; its explosive popularity surprised even China.

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

For Yang Zhilin, this was a double kill. The arbitration was unresolved, and R1 directly sentenced his past year's strategy to death: users acquired through spending were insignificant in front of a free and stronger competitor. Public opinion turned its guns, with one article titled "Yang Zhilin, a Suspended Idealist of the Post-90s Generation." Suspended means not grounded. He worried about becoming utilitarian while writing songs, and now the whole world says he is both utilitarian and a failure.

At the beginning of 2025, the Dark Side of the Moon still had money in the account, but very little say.

The Turnaround Against Gravity

The next year was crucial for Yang Zhilin.

Yang Zhilin almost completely denied his past year. He stopped advertising, cut redundant businesses, shrank to the basic model, and turned to open source. In a discussion at Geek Park, he said that organizational inertia wants to do more and more things, "we need to fight against this gravity."

This sounds easy, but doing it means admitting the route was wrong, laying off people he hired, and bowing to the competitor that nearly killed him. Most 33-year-old founders cannot overcome this hurdle. Yang Zhilin did it exceptionally decisively, perhaps because open source and long-termism were his original settings, while closed-source user acquisition was merely a garment put on by capital, which he has now taken off.

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

In the early hours after the release, Yang Zhilin, along with Zhou Xinyu and Wu Yuxin, held an AMA on Reddit, answering 21 questions. He clarified that the $4.6 million training cost was not the official figure, admitting that the number of GPUs was not as many as their American counterparts, "but we have squeezed the performance of each card to the limit." When asked about OpenAI's cash burn, Zhou Xinyu replied relaxedly: We don't know, only Sam knows, "we have our own rhythm."

Their own rhythm. The Dark Side of the Moon in 2024 could not articulate these five words; at that time, its rhythm was dictated by investors and the ROI of advertising. The one who returned these five words to Yang Zhilin was precisely Liang Wenfeng. R1 proved that open source combined with algorithmic efficiency works in China, effectively giving Yang Zhilin a roadshow to his board. The person who was nearly killed by DeepSeek now has to thank it for his life.

The market's return was also very direct. On the last day of 2025, the Dark Side of the Moon officially announced a $500 million Series C round, with Alibaba, Tencent, and Wang Huiwen all increasing their investments, valuing the company at $4.3 billion, with cash exceeding 10 billion. Less than 20 days after the release of K2.5, revenue surpassed the entire year of 2025, and personal subscription orders increased more than eighty times month-on-month, breaking into the top ten of Stripe's global rankings. Then came this week's K3.

The PhD student who seven years ago said, "not trying will lead to lifelong regret," has now made the American tech circle start to question life, becoming an example used to hit the White House.

The Hermit's Bill

But reality never only slaps one side of the face. After 2025, it was Liang Wenfeng's turn to pay the bill.

He has no life, but his employees do. Luo Fuli, Wang Bingxuan, Wei Haoran, Ruan Chong, these names are core backbones within DeepSeek, and starting in 2025, they gradually left, with many taking on leadership roles elsewhere.

The street talk is quite heartbreaking: when colleagues of similar levels can jump out and earn so much, why can't I? A research utopia without levels, KPIs, or discussions about money relies on members' loyalty to the problems themselves. But R1 raised everyone's market value tenfold, and loyalty now has a priced competitor for the first time.

Money is also starting to have problems. The management scale of Huanshuo shrank from its peak to over 20 billion, with the battery powered by love leaking electricity. Thus, an unimaginable scene emerged: the person who once said, "there are no short-term financing plans" began to meet with investors.

He initially offered a minimum investment of 5 billion, later lowered to 1.5 billion. But throughout the negotiations, he repeatedly emphasized not valuation, not shareholding ratio, but one condition: no poaching of DeepSeek's people, no inciting them to start their own businesses. A financing deal was struck with a non-poaching agreement. He could give up shares, but not the atmosphere of that laboratory.

As the two lines progressed into 2026, an unexpected situation emerged.

Yang Zhilin, while cutting advertising, squeezing the performance of every H800, and discussing his own rhythm, increasingly resembled Liang Wenfeng. Liang Wenfeng, meanwhile, was meeting VCs, worrying about retaining talent, and for the first time being distracted by such worldly issues, having to step out of the laboratory.

One was accumulating water, the other was seizing the tide; now the one accumulating water found that the reservoir could also leak, while the one seizing the tide finally awaited his own wave.

Two Guangdong Men Rewrite the Chinese AI Table

This is the most bustling time for AI entrepreneurship in China, with two names pushed to the forefront: Liang Wenfeng and Yang Zhilin.

They do not resemble the founders familiar to the Chinese internet over the past decade. There are no quotable posters, no legendary dinner tales, and no strong desire to shape themselves into legendary entrepreneurs. Yet their positions are closer to the intersection of money, power, and the emotional zeitgeist of the era than most entrepreneurs in the past.

What sets Liang Wenfeng apart is that he appears to have almost no "life." In public reports, he seems more like a person consumed by work: from quantitative investment to building computing clusters, and then to DeepSeek, his narrative rarely includes family, consumption, hobbies, or socializing, almost entirely filled with models, architectures, computing power, organization, and originality.

What sets Yang Zhilin apart is that he seems to have no "retreat." After the establishment of the Dark Side of the Moon, Kimi quickly became one of the most watched AI applications in China, with financing, valuation, user growth, model iteration, commercialization, and arbitration from old shareholders all pressing down on this company simultaneously. The faster it runs, the less it can afford to stop and explain for too long.

This is where the two are most alike and yet most different.

Liang Wenfeng seems like a technical idealist emerging from deep within the capital market. He first proved that AI could directly change the flow of money in Huanshuo Quantitative, and then turned this capability toward large models. The story of DeepSeek did not initially rely on financing but on the funds, computing power, and engineering culture accumulated from Huanshuo. What the outside world later remembered were R1, V3, low-cost training, and open-source shaking the American market, but more crucially: Liang Wenfeng defined the problem early on as the gap between "originality and imitation," rather than "how to make money from Chinese large model applications."

This makes him appear anti-commercial yet extremely commercial.

Anti-commercial in that he repeatedly downplays short-term monetization in public interviews, unwilling to portray DeepSeek as a story of user acquisition and market share grabbing. Extremely commercial in that every step he takes directly targets the cost structure at the very bottom of the AI industry: training efficiency, reasoning costs, model architecture, talent density, and chip constraints. If a model can approach top capabilities at a lower cost, it changes not just a product ranking but the entire industry's imagination regarding capital expenditure.

This is also why, when DeepSeek exploded in popularity, the global market's reaction was so intense. It was not just another Chinese chatbot; it reminded investors that the most expensive logic chain in the American AI narrative over the past two years might not be as solid as imagined. Larger models, more GPUs, and higher capital expenditures do not necessarily equate to an insurmountable moat.

Yang Zhilin faces a different fate.

From its inception, the Dark Side of the Moon has stood in the spotlight. The founder's resume is impressive enough: Tsinghua undergraduate, Carnegie Mellon PhD, involved in significant research like Transformer-XL and XLNet; the company was founded in 2023, and Kimi differentiated itself with long context, quickly becoming an AI product that ordinary users could perceive. It was not discovered by the tech circle first as DeepSeek was, but entered the public product, capital market, and industry narrative much earlier.

This gave Yang Zhilin a tremendous advantage, but also a huge burden.

The advantage is that Kimi has a product mindset. Many people first seriously use domestic AI not because they understood a technical report, but because Kimi can read long texts, organize materials, and handle workflows. AI is moving from the laboratory to the office, and there needs to be an entry point that ordinary people can remember; Kimi once seized this position.

The burden is that once a product mindset is established, it must be continuously nurtured. Users will wait for stronger models, investors will wait for higher revenues, teams will wait for greater option values, and competitors will wait for you to make mistakes. The more financing "The Dark Side of the Moon" receives, the higher its valuation, and the less likely Yang Zhilin is to return to a quiet researcher position.

By 2026, this pressure will become clearer. Public reports indicate that "The Dark Side of the Moon" will complete approximately $2 billion in new financing in May 2026, with a post-investment valuation exceeding $20 billion; the company's official website has also placed Kimi K2, code, and Agent capabilities in a core position. The capital market does not give it applause, but a bill: you must prove that you are not just "the best product-making large model company," but also that you can maintain dual leadership in technology and business amidst the encirclement of DeepSeek, Alibaba, ByteDance, MiniMax, and Zhizhu.

What’s more troublesome is that Yang Zhilin is shadowed by old entrepreneurial projects. In 2024, media such as the South China Morning Post reported that Yang Zhilin, founder of "The Dark Side of the Moon," and co-founder Zhang Yutao were arbitrated by some previous investors of Loop Intelligence in Hong Kong; Yang Zhilin's side claimed that the necessary procedures for leaving Loop Intelligence and starting a new venture had been completed, while the investors presented a different narrative. The core of this dispute is not just a personal conflict of a certain founder, but a sharper question within the wave of AI entrepreneurship in China: when the value of new companies skyrockets, where exactly are the boundaries between old companies, old shareholders, old teams, old intellectual property, and new financing?

Yang Zhilin cannot just portray himself as a genius researcher. The scale of financing, commercialization revenue, product iteration, IPO expectations, and legal disputes will require him to become a more complete and ruthless CEO. A researcher can prove themselves with papers, but a CEO must prove themselves through the organization. Papers can have names, but organizations cannot operate solely on signatures.

The more low-key a person is, the more they are magnified by the times. The more a person wants to move forward, the more they are chased by the past.

This round of stories in Chinese AI may not ultimately be won by the best storytellers. It is more likely to be won by those who can simultaneously endure three things: the uncertainty of technology, the patience consumption of capital, and the cost of founders being both mythologized and judged.

Whether Liang Wenfeng has a life, the outside world actually does not know. Whether Yang Zhilin has a way out has not yet reached a final answer.

But at least for now, they both have little space to return to the status of ordinary people.

Original link