Wallstreetcn Reviews Silicon Valley's Chinese AI Entrepreneur Community: From HRT Interns to Big Tech's Core

On October 19, 2026, Alexandr Wang posted a photo of old friends on the social platform X.

In the photo, Wang, the head of Meta Superintelligence Labs, Scott Wu, founder of the AI coding company Cognition, and Jeff Yan, founder of the crypto trading platform Hyperliquid, were reunited. Wang's caption was just one line: "got the HRT gang back together!"

This was a reunion spanning a decade. About ten years ago, these young people, who had stood out from math and science competitions and top universities, studied high-frequency algorithms together in the internship office of the quantitative giant Hudson River Trading (HRT); ten years later, they have moved on to different commercial heights—Wang founded Scale AI and then joined Meta to build the star project Muse, Wu is trying to get AI to write software independently, and Yan has built a star derivatives platform in crypto finance. The interns of that time have each grown into key kingmakers who influence the direction of the industry.

And in today's Silicon Valley, this photo is taking on another meaning: the same cohort of talent is entering the same technological revolution at an astonishing density, and rapidly reaching the center of capital, technology, and commercial power.

Silicon Valley once had its own "PayPal Mafia"—a group of entrepreneurs who came out of the same company and later jointly shaped the internet era.

Image source social media account: AYi

At the most fundamental layer of computing infrastructure, the world's four major semiconductor giants are led by people of Chinese descent—Nvidia's Jensen Huang, AMD's Lisa Su, Broadcom's Hock Tan, and Intel's Lip-Bu Tan—who firmly control the supply lifeline of global GPUs and custom AI chips.

At the middle layer of the data supply chain and frontier labs, Alexandr Wang of Scale AI and Edwin Chen of Surge AI split the core market for high-quality training data; Meta Superintelligence Labs has Zhao Shengjia as chief scientist, while his rival OpenAI has Mark Chen leading the overall R&D effort, and xAI's founding technical team also includes Chinese core members such as Jimmy Ba and Yuhuai Wu.

At the top layer of commercial applications, in addition to Scott Wu's Cognition, Johnny Ho has pushed Perplexity's valuation to tens of billions of dollars and reshaped the search experience, Jesse Zhang's Decagon has captured a multibillion-dollar customer service market, Qiao Lin's Fireworks AI has become an important hub for open-source model inference, and 24-year-old math prodigy Carina Hong is taking on machine theorem proving with Axiom Math.

To understand where this density of talent comes from, it helps to first step into an ordinary house where young entrepreneurs live.

In the fall of 2025, a Rest of World reporter moved into the old residence where Mark Zuckerberg lived in his early years. This light blue bungalow in Los Altos, Silicon Valley, is known as the "Facebook House." Twenty years ago, Zuckerberg worked here and later built a social networking empire; today, it has become a foothold for a new generation of Chinese technical talent.

Screenshot of Rest of World's report The Chinese whiz kids of Silicon Valley

The house is simply furnished, with mattresses, bookshelves, and a few cats, yet it hosts high-density AI discussions year-round. At one gathering, more than thirty people sat around talking, a whole sheep was roasting in the yard, and the topics ranged from embodied intelligence and social products all the way to large model architecture. A robotics product manager who attended the gathering sighed that the atmosphere around them had become "if you don't do AI, it's as if you don't deserve to live."

This scene reveals the actual operating mechanism of the so-called "Chinese gang."

It has no unified command center, nor any secret alliance of interests. It is first and foremost a talent network woven from a common language, competition records, overseas study experiences, and technical communities. Here, newcomers to the Bay Area look for housing, engineers exchange testing details of frontier papers, and entrepreneurs look for technical co-founders. Recruitment information and investment leads spread through overlapping WeChat groups and the social platform X.

Catalyzed by capital, this densely woven talent network quickly made its way onto the AI business table. Scott Wu, Alexandr Wang, Edwin Chen, and Carina Hong—this group of tech geeks who graduated from top universities such as MIT, Harvard, and Stanford, hold gold medals in math/informatics olympiads, and mostly honed their skills at quantitative giants like HRT—entered the same wave.

Scott Wu is betting that AI can truly take on software development work. In March 2024, his company Cognition launched Devin, attempting to evolve AI from a code completion tool into an engineer that can take on complete development tasks. The product launch caused a sensation and also drew skepticism: can the capabilities shown in the demo be reliably delivered in real projects? To bridge the gap between demo and reality, Wu chose to go deep into enterprise delivery and positioned the product as a cross-model orchestration architecture:

The world spends trillions of dollars a year on "meat computers"—that is, human software engineers... We quite enjoy being "Switzerland." If you do everything, you'll lose to Microsoft or Google. They do everything, have trillions more dollars in resources than you, and a hundred thousand more people than you.

Video screenshot, same below

In 2026, Cognition announced it had raised more than $1 billion in a large funding round, reaching a post-money valuation of $26 billion, with annualized revenue run rate (ARR) climbing to $492 million, and its valuation then surged to $48 billion, adding Goldman Sachs, Mercedes-Benz and others to its client list. But the "neutral Switzerland" model soon ran into challenges: six days before the group photo was posted, Meta, led by Alexandr Wang, released Muse Wonder and ordered that the most powerful version, Muse Wonder Full, be built only into its in-house Muse Code, directly cutting off external access. The giant itself entered the application layer, squeezing the room for independent orchestration to survive.

Alexandr Wang's fate moved toward "institutionalization." The Scale AI he founded had once been the industry-wide data infrastructure. In June 2025, Meta announced an investment of about $14.3 billion in Scale AI, valuing the company at more than $29 billion, and Wang subsequently joined Meta to lead its superintelligence lab.

The deal brought Wang into the power center of a giant, but it dismantled the independence on which Scale AI depended for survival. Out of concern for preventing leaks of core model assets, OpenAI and Google quickly adjusted their cooperation, and Reuters later disclosed that Google planned to cut an annual procurement contract of about $200 million. For a data service provider nurtured by all the major companies together, losing neutrality is itself the biggest commercial damage.

Edwin Chen chose a different kind of bet. A graduate of MIT with degrees in computer science, mathematics and linguistics, he had been responsible for algorithm engineering at Twitter and Google. When he founded Surge AI, he insisted on an independent model with zero outside venture capital: "VC makes companies bloated and distorts a team's incentives." According to data disclosed by The Information and Sacra, Surge AI achieved $1 billion to $1.2 billion in revenue in 2024 with about 110 full-time employees, surpassing Scale AI, which had a team of a thousand people in the same period (about $870 million), and remained consistently profitable, naturally taking on the data orders flowing out from the major companies.

In the deeper field of mathematical verification, 24-year-old Carina Hong dropped out of a doctoral program at Stanford to found the theorem-proving AI company Axiom Math, and in March 2026 completed a $200 million Series A at a post-money valuation of more than $1.6 billion. She recruited the renowned mathematician Ken Ono to her team, focusing on using formal verification to solve the hallucination problem of large models in complex reasoning. In an interview, she clearly defined the technical route:

Four people, four different fates: Wu wants to turn a technical sensation into a reliable product, Wang has gone from entrepreneur into the power center of a giant, Chen seeks to win the market through independence, and Hong is looking for a commercial outlet at the frontier of mathematical reasoning.

From enterprise programming and neutral data to theorem verification, they quickly cut into the frontier with extremely high technical judgment, but under the reshuffling of capital and the giants, former acquaintances also have to face harsh market competition.

Behind the rapid advancement of technical products, the competition among frontier labs for core R&D talent has also become white-hot.

In mid-2025, Zuckerberg personally stepped in and successively poached four core researchers from OpenAI: Zhao Shengjia, Yu Jiahui, Ren Hongyu and Bi Shuchao.

All four technical backbones completed their undergraduate education at prestigious domestic universities (Tsinghua, the USTC Special Class for the Gifted Young, Peking University and Zhejiang University), then obtained doctorates at Stanford, UIUC and Berkeley. Among them, 29-year-old Zhao Shengjia was deeply involved in the architecture design of ChatGPT, GPT-4 and the o1 logical reasoning model, and after joining was appointed chief scientist of Meta's superintelligence lab. Zuckerberg publicly stated:

Zhao Shengjia has already pioneered several breakthroughs, including new scaling paradigms, proving himself a leader in the field.

Mark Chen, senior vice president of research at OpenAI responsible for overall technical defense, worked at Jane Street in his early years and once served as a coach of the U.S. informatics national team together with Scott Wu. Zuckerberg once privately invited him to join, but Mark Chen refused.

The answer cannot be reduced simply to "being good at math" or "working harder." Looking at these people together, their growth paths are not entirely the same, yet there are some recurring intersections: training in mathematics and computer science, top universities, research institutions, quantitative trading or large tech companies, and Silicon Valley's startup networks. Rigorous math and science olympiad training in basic education laid the underlying thinking, top North American universities built the core technical and personal networks, and experience at quantitative institutions such as HRT and at tech giants sharpened a keen intuition for ultra-high-concurrency architectures and extremely complex systems. This path directly converts extreme mathematical talent into hardcore engineering execution, becoming the most certain incubator in this round of AI explosion.

For example, Scott Wu's background is tied to programming and math competitions; Alexandr Wang is known for his strength in mathematics; and Carina Hong also moved from an interest in math and intensive technical education into AI math entrepreneurship. This is not a path unique to ethnic Chinese, but it offers a clue for understanding the current concentration of talent.

Early Chinese and Indian immigrants had also accumulated decades of technical and entrepreneurial experience in Silicon Valley. A 1999 study by UC Berkeley scholar AnnaLee Saxenian found that in 1998, Chinese and Indian engineers already held leadership roles at about a quarter of Silicon Valley's high-tech companies. Today's AI startup wave did not start from zero; it is built on talent flows and professional networks formed over the long term.

A global AI talent tracking report titled "Who's Ahead in the Global AI Talent Race?", released by the Carnegie Endowment's Carnegie China in September 2026, shows that among the world's top AI researchers, as many as 57% completed their undergraduate education at Chinese universities, a jump of 11 percentage points from 46% in 2022; over the same period, the share of talent trained at the undergraduate level in the United States fell to 13%. The study wrote:

China has established a significant lead as the largest source country of global elite AI talent.

The Carnegie study used a sample of 10,280 NeurIPS paper authors.

Industry has likewise felt this depth of talent. Nvidia CEO Jensen Huang said bluntly in an interview with Axios that "half of the world's AI researchers are Chinese."

The appeal of the U.S. tech industry has given some of this talent the opportunity to turn research capabilities into entrepreneurial opportunities. Investors, early employees and potential customers can connect with one another through shared schools, former employers and industry gatherings. Spaces like "Facebook House" are a microcosm of this network in everyday life.

However, this does not mean that ethnic Chinese hold comprehensive dominance in business. But looking at this wave of generative AI, the presence of other ethnic groups is indeed far less dense than that of ethnic Chinese. Most Western elites remain in the capital-and-vision narrative of the previous wave, while people of Indian descent have mostly taken the route of big-company CEO—Satya Nadella and Sundar Pichai are both examples—rather than the route of startup leader or chief scientist. Eastern European and Russian scientists are still making breakthroughs in theoretical exploration, but it is ethnic Chinese who are emerging in large numbers as entrepreneurs and core R&D personnel.

Returning to the cases of Scott Wu and Carina Hong, this technology cycle has profoundly changed the path by which technical talent enters business.

Ten years ago, developing complex software required hiring a large workforce to build toolchains and infrastructure, and it took a long development process before reaching the market. Now, underlying large models output general reasoning capabilities through APIs, cloud providers offer ready-made computing foundations, and startup teams can focus entirely on orchestration optimization, prompt engineering and specialized delivery.

A small team can develop a coding agent or a theorem-verification system without building its own underlying model. This greatly lowers the threshold for launching technical prototypes, enabling small engineering teams with strong math and code-optimization capabilities to launch products that support enterprise procurement within months.

But this speed is also a double-edged sword: the application layer is easy to enter, and giants can follow just as quickly. Once a product is absorbed by an upstream model platform, independent startups must withstand enormous pressure to survive. A technical breakthrough only gets you a ticket to the game; how not to be swallowed by the platform is the long-term question.

Two generations of ethnic Chinese, two paths

Looking only at young entrepreneurs, it is easy to mistakenly think that the rise of ethnic Chinese in Silicon Valley began with generative AI. In fact, older-generation ethnic Chinese leaders such as Jensen Huang and Lisa Su had long since built hardware barriers in the chip industry. Jensen Huang runs Nvidia, Lisa Su leads AMD, Hock Tan heads Broadcom, and veteran venture investor Lip-Bu Tan serves as Intel CEO.

Photo collage of Jensen Huang, Lisa Su and Lip-Bu Tan at public events

But the paths of the two generations are completely different: the older generation built industrial power through decades of manufacturing, supply chains and ecosystems, while the new generation, on top of ready-made computing power, quickly enters the application layer with extremely high capital efficiency. The vast majority of the huge sums raised by the younger generation ultimately flow to the data centers of Nvidia, AMD and Broadcom—software speed is always dependent on the hardware foundation.

In Silicon Valley's history, immigrants have always been an important driving force. In different eras, talent from different regions entered core positions through education, research, engineering and entrepreneurship.

And in this round of the AI gold rush, those surging to the forefront in large numbers are ethnic Chinese. They are not one single "gang," yet they share similar training, networks and ambitions. What is truly worth asking is not how they compete, but how this network of ethnic Chinese talent will define the next stage of Silicon Valley. Risk disclosure and disclaimer

Markets carry risk, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation or needs of individual users. Users should consider whether any opinions, views or conclusions in this article are suitable for their particular circumstances. Investment based on this article is at the user's own risk.