Understanding the Eight Leading AI Companies in China and the US, and the Founders Behind the Models

Compiled by: PANews

There is no single correct answer to AI entrepreneurship. Researchers with a focus on papers, academic commercialization teams, quantitative engineers, cross-industry entrepreneurs, and operational organizers can all find their place in this competition.

Over the past decade, artificial intelligence has transformed from cutting-edge research in laboratories to a core force shaping the technology industry landscape. Today, the most watched AI companies globally include OpenAI, Google DeepMind, Anthropic, xAI, as well as DeepSeek, Dark Side of the Moon, Zhiyu AI, and MiniMax.

On the surface, they are all developing large models, but they represent completely different entrepreneurial paths: some come from top papers and academic research, some push academic research results into industry, some accumulate computing power and capital in quantitative finance, and others excel in financing, products, operations, and organization, turning technical teams into world-class companies.

🇨🇳 China: From Research Papers and Academic Laboratories to Quantitative Engineering

Dark Side of the Moon: Turning a Game-Changing Paper into a Product for Everyone

Founded in 2023, Dark Side of the Moon's product Kimi is one of the most recognizable applications of large models in China, with the Kimi K3 model focusing on multi-modal, long context, programming, and deep reasoning.

Founder Yang Zhilin, born in 1992 in Shantou, Guangdong, graduated from Tsinghua University and later obtained a PhD in Computer Science from Carnegie Mellon University, studying under researchers in natural language processing. During his doctoral studies, he participated in widely cited works such as Transformer-XL and XLNet, and has also conducted AI research at Google Brain and Meta.

Among this group of Chinese entrepreneurs, Yang Zhilin has taken the most typical path—first proving his judgment in the academic circle, then transforming that judgment into a product that ordinary people can directly use. Compared to Zhiyu, which leans more towards enterprise services, Dark Side of the Moon has aimed at the consumer end from the very beginning.

Zhiyu AI: A Model of Industrialization Emerging from Tsinghua's Laboratory

Founded in 2019, Zhiyu AI originates from Tsinghua University's Knowledge Engineering Laboratory system, with its core technology route being the GLM series, focusing on foundational models, enterprise services, and model platforms.

Co-founder and CEO Zhang Peng has studied at Tsinghua University for his bachelor's, master's, and doctoral degrees, and has been deeply involved in the Knowledge Engineering Laboratory and AMiner academic platform construction from 2005 to 2020, without a background in traditional internet giants.

What sets Zhiyu AI apart is its starting point—it did not begin from a hit product but directly transformed over a decade of accumulated knowledge graphs and academic networks from a university laboratory into a foundational model platform that can provide services externally. This is also the biggest distinction between it and Dark Side of the Moon: one started with a platform, while the other started with a product.

DeepSeek: Investing Profits from Quantitative Trading into Large Models

Founded in 2023, DeepSeek focuses on foundational models, training efficiency, and open weights, with representative models including DeepSeek-V3, R1, and the upcoming V4 series.

Founder Liang Wenfeng was born in 1985 in Wuchuan, Guangdong. He studied Electronic Information Engineering at Zhejiang University and later obtained a master's degree in Information and Communication Engineering, with research interests in machine vision. He does not have a background in major companies like Huawei, Tencent, or Baidu, but instead immersed himself in machine learning quantitative trading during his graduate studies and later co-founded Huansheng Quantitative.

Liang Wenfeng's path is the most unique among these companies: he did not start as a researcher looking for funding but first accumulated capital and computing power through quantitative investment, and then invested these resources into large model research. In a sense, DeepSeek is a product of "computing power earned from finance, fed to AI."

MiniMax: From Senior Technical Executive in Big Companies to Multi-Modal Entrepreneur

Founded in 2022, MiniMax is a multi-modal foundational model company covering text, speech, images, and video, with products including Conch AI and Xingye.

Founder Yan Junjie holds a bachelor's degree in mathematics from Southeast University and a PhD in artificial intelligence from the Institute of Automation, Chinese Academy of Sciences, and completed postdoctoral research at Tsinghua University. Before starting his own business, he worked at SenseTime for over six years, rising to the position of Vice President and Deputy Dean of the Research Institute.

Yan Junjie embodies two types of experience—solid academic training and practical experience in managing products and teams within a large-scale AI team like SenseTime. MiniMax's choice to pursue the more challenging multi-modal path is, to some extent, a natural choice resulting from the combination of these two experiences.

🇺🇸 United States: From Basic Research and Safety Governance to Cross-Industry Resource Integration

Anthropic: A Sibling Duo, Half Physicist, Half Governance Expert

Founded in 2021 by a group of researchers who left OpenAI, Anthropic's core philosophy is that "AI safety should advance in tandem with capability development," with its representative product being the Claude series.

Co-founder and CEO Dario Amodei studied physics at Stanford University for his undergraduate degree and shifted to biophysics at Princeton University for his PhD. This interdisciplinary background later influenced his research path—he first joined Google Brain as a senior research scientist and then moved to OpenAI as vice president of research, participating in the training of GPT-2 and GPT-3, and was one of the core participants in early RLHF (Reinforcement Learning from Human Feedback) research. In 2021, he left OpenAI with several colleagues, including his sister Daniela, to establish Anthropic.

Daniela's path is entirely different. She studied English literature for her undergraduate degree and did not enter the tech industry directly after graduation, instead working in operations at Stripe and OpenAI. At Anthropic, her role is not to train models but to build the foundational systems that allow the company to operate safely and stably—from recruitment and corporate governance to the implementation of safety policies.

This sibling duo, to some extent, embodies Anthropic itself: it requires researchers like Dario who can judge technical routes and understand model capability boundaries, as well as individuals like Daniela who comprehend organization, processes, and risk management.

OpenAI: A Storyteller and a Technical Expert

Founded in 2015, OpenAI has launched iconic products like ChatGPT, GPT, and Sora, and is a core company that brought generative AI into the public eye.

Sam Altman dropped out after two years of studying computer science at Stanford University to start a location-based social networking company, Loopt, and later became president of Y Combinator. His strengths have never been in publishing top papers but in financing, product judgment, organizational mobilization, and telling a compelling technical narrative.

Ilya Sutskever, on the other hand, represents a different type. He holds a PhD in mathematics and computer science from the University of Toronto, studying under deep learning pioneer Geoffrey Hinton, and participated in significant achievements like AlexNet and Seq2Seq at Google Brain, leading OpenAI's early large-scale training route.

If Ilya determines what kind of models OpenAI can create, then Sam decides whether the company can survive, secure funding, and sell its technology. The combination of these two individuals is perhaps the most successful practice of the "scientist + organizer" model in Silicon Valley.

xAI: Musk Transferring His Skills from Cars and Rockets to AI

Founded in 2023, xAI's representative product is Grok, emphasizing real-time information, reasoning, and generative media, deeply integrated with the X platform.

Founder Elon Musk was born in Pretoria, South Africa, studied physics and economics in college, and briefly entered a graduate program at Stanford before dropping out. Unlike other AI founders, he does not have a traditional corporate background but has been involved in creating companies like Zip2, PayPal, SpaceX, and Neuralink, while also leading Tesla for a long time.

Musk's entry into AI did not start from papers or laboratories but from integrating resources accumulated across industries such as payments, electric vehicles, aerospace, satellite communications, and social platforms, directly investing them into the computing power and data needed for large model training.

Google DeepMind: A Chess Genius Taking a Path from Games to Proteins

Founded in 2010 and acquired by Google in 2014, Google DeepMind is now Google's core AI research institution, having launched AlphaGo, AlphaFold, and Gemini.

Co-founder and CEO Demis Hassabis was born in London, England, holds a first-class degree in computer science from Cambridge University and a PhD in cognitive neuroscience from University College London. He was a chess prodigy in his youth, participated in the development of the game "Theme Park" at 17, and later founded his own game company, Elixir Studios.

His gaming experience trained his intuition for complex systems and decision-making, neuroscience provided him with a biological perspective on understanding "intelligence," and computer science equipped him with the engineering capabilities to realize these ideas. The success of AlphaGo and AlphaFold is almost the result of the combination of these three experiences—the latter also earned him the Nobel Prize in Chemistry in 2024.

Four Founders' Paths: There is No Single Template for AI Entrepreneurship

Looking back at these eight companies, the founders' backgrounds can be roughly divided into four categories.

The first category: purely technical paper-oriented founders. Yang Zhilin, Yan Junjie, Dario Amodei, Ilya Sutskever, and Demis Hassabis all have systematic research training, sharing the commonality of being able to judge technical routes, lead research teams, and form real barriers in model capabilities or training methods. Among them, Yang Zhilin and Ilya are closer to the path of "first writing game-changing papers, then moving towards entrepreneurship"; Hassabis is a product of the intersection of computer science, gaming, and neuroscience.

The second category: industrialization of academic research results. Zhiyu AI is the most typical example, with core capabilities directly derived from years of accumulation in Tsinghua's Knowledge Engineering Laboratory. Dark Side of the Moon also has a strong Tsinghua and overseas research gene, but its path is more market-oriented—it did not stop at packaging academic results into enterprise services but directly created a product for the public.

The third category: a combination of mathematics, engineering, and capital. Liang Wenfeng and Musk belong to this category, but their sources of resources are completely different—Liang Wenfeng accumulated capital and computing power through quantitative investment, while Musk mobilized resources from cross-industry businesses such as payments, automotive, aerospace, and social platforms. The commonality is that their entry into AI has never relied solely on algorithms.

The fourth category: non-technical organizers paired with technical founders. Sam Altman and Daniela Amodei are representatives of this route. They may not be the top model researchers, but they determine whether the company can recruit talent, secure funding, build products, withstand regulatory pressure, and not fall apart during rapid expansion.

The competition in cutting-edge AI has never been just about "who has a PhD in computer science" or "who can train larger models."

What truly determines whether a company can maintain its lead is whether the capabilities of research, engineering, products, capital, and organization can form a closed loop: scientists determine the technical ceiling, engineering teams decide whether models can be implemented, product teams determine whether users are willing to use them, capital determines the investment in computing power and talent, and organization and governance determine whether the company can operate sustainably.

From this perspective, there is no single correct answer to AI entrepreneurship.

Researchers focused on papers, academic commercialization teams, quantitative engineers, cross-industry entrepreneurs, and operational organizers can all find their place in this competition.

Those who can integrate different capabilities into a system are more likely to become long-term winners in the next wave of artificial intelligence.

Disclaimer: This content is provided for general branding and informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online events, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets or to use any services. Crypto assets are highly volatile and may result in loss. WEEX services and online events may not be available in all regions and are subject to applicable laws, regulations, and eligibility requirements. You are responsible for ensuring that your use of WEEX services complies with local laws and for carefully assessing the risks before participating in any crypto-related activities.