The secret of a 28-year-old Chinese-American genius billionaire: dropping out of school to start a business due to a stolen refrigerator, wealth skyrocketing with "AI laborers," and being poached at a high price by Silicon Valley tycoons

On June 24, ZhDongxi reported that on June 13, the rumors of a sensational deal in the AI industry finally settled. After more than two months of negotiations, this investment raised the valuation of Scale AI to $29 billion (approximately RMB 208.2 billion). Meta ultimately acquired 49% of the shares of the well-known AI data unicorn Scale AI for $14.3 billion (approximately RMB 102.8 billion) and brought Alexandr Wang, the 28-year-old co-founder and CEO of Scale AI, under its wing. In recent days, several foreign media outlets have intensively disclosed more negotiation details of this significant deal, such as Alexandr Wang's new position at Meta, the additional "control clauses" proposed by Meta regarding Scale AI and Alexandr Wang, and have deeply explored Alexandr Wang's growth path and the founding history of Scale AI. It is reported that Zuckerberg reached out to Alexandr Wang as early as April, wanting to recruit him to Meta, initially offering $5 billion (approximately RMB 35.9 billion). After nearly two months of negotiations, they reached a final agreement of $14.3 billion (approximately RMB 102.8 billion). Based on the shareholding ratio, Alexandr Wang himself will benefit over $1 billion (approximately RMB 7.2 billion) from this deal. Although Alexandr Wang will step down as CEO, he will remain on the board of Scale AI and may serve as an executive in the newly established "Superintelligence" department within Meta. There are also reports discussing the possibility of him becoming Meta's Chief AI Officer. As a star entrepreneur in the AI industry, Alexandr Wang shares many similarities with Zuckerberg, both having started their ventures during their undergraduate studies, both being dropouts from top American universities, and both being embroiled in intense controversies. With the sweat of hundreds of thousands of outsourced data labeling workers around the globe, Alexandr Wang's wealth has rapidly expanded alongside the development of the AI industry, giving rise to one of the youngest self-made billionaires in the world and the youngest female billionaire—Lucy Guo, the co-founder of Scale AI who dropped out midway. His network is also widely discussed, not only having a close friendship with Sam Altman, co-founder and CEO of OpenAI, but also extensively networking with politicians, with a social intensity that is astonishing; it is said that "there is no one in the AI circle he doesn't know." ▲Co-founder and CEO of data labeling startup Scale AI, Alexandr Wang (Image source: Time Magazine) From the transaction documents, in addition to equity, Meta will also strengthen its control over Scale through mechanisms such as "poison pill" and "liquidation preference." Alexandr Wang himself faces a dual challenge: he must meet Meta's expectations for breakthroughs in AI technology while maintaining Scale AI's position as a neutral data service provider without overly involving himself in Scale AI's business. Scale AI's clients include direct competitors of Meta, such as OpenAI and Google. Meta's entry has raised concerns among these clients about data security and the potential leakage of commercial secrets, leading them to begin scaling back their cooperation. As Scale AI's largest client, Google originally planned to pay approximately $200 million (about 1.44 billion RMB) this year for manually labeled training data, which is crucial for developing cutting-edge AI models, including the Gemini chatbot. After learning this news, Google's parent company Alphabet plans to terminate its partnership with Scale AI. It is reported that Google has begun discussions with several of Scale AI's competitors to transfer most of its data labeling business. Why is Meta going to such lengths to bring Alexandr Wang under its wing, and why is it making such a significant investment in Scale AI? How will Meta, having just experienced an antitrust storm, persuade regulators to approve this deal? How will Scale AI, without its founder, navigate this turbulent period? This article will deeply review the behind-the-scenes details and the ins and outs of this massive AI transaction, as well as revisit the legendary entrepreneurial journey of Alexandr Wang, a 90s-born Chinese-American genius, over the past nine years. Zuckerberg Started Recruiting in April, Negotiation Prices Soared In mid-April, Meta founder and CEO Mark Zuckerberg reached out to Alexandr Wang, expressing a desire for him to join Meta and offering to pay billions of dollars to facilitate this. According to transaction documents and sources familiar with the matter, Zuckerberg proposed an investment of $5 billion (about 35.9 billion RMB) in Scale in early May, to which Alexandr Wang subsequently raised the price to $20 billion (about 143.6 billion RMB). By early June, negotiations stalled as the Scale board engaged in heated debates over how much control to cede to Meta. Although Alexandr Wang holds super voting rights through a special share structure, unlike Zuckerberg's absolute control over Meta, he does not possess a majority of voting rights in Scale. The risks of cooperation are evident: aligning with Meta could lead to Scale losing orders from competitors like Alphabet and OpenAI Top law firms in Silicon Valley have intervened in the negotiations, with Latham & Watkins representing Meta and Wilson Sonsini advising Scale, which has also hired investment bank Centerview Partners for assistance. Among the four-member board of Scale, only Dan Levine from Accel and former Index Ventures partner Mike Volpi, both early investors, are involved in decision-making, while the other two are Alexandr Wang and his friend, William Hockey, co-founder of fintech company Plaid. The board had considered enforcing a "buy-sell option" to compel Meta to acquire the remaining shares in the future, but ultimately abandoned this idea to avoid a breakdown in negotiations. Meta executive Dave Wehner made it clear to Volpi that if they could not gain control over the investment, Meta would not accept a deal to acquire a 49% stake, and they were even less willing to be forced to sell shares as minority shareholders. On the evening of June 5, the Scale board finally approved the deal. To stabilize morale, the company set aside hundreds of millions of dollars as a bonus pool for employees with unvested options, while former employees who had not exercised their options would need to liquidate their options through a tender offer. On June 11, the board officially recommended that shareholders accept Meta's acquisition, and on June 13, Alexandr Wang tearfully bid farewell to employees at the San Francisco headquarters, receiving a standing ovation. The final transaction price was close to Alexandr Wang's initial asking price, with Meta acquiring a 49% stake in Scale for $14.3 billion (approximately RMB 102.8 billion), slightly above the valuation from the startup's financing round a year ago. It is reported that this young billionaire will simultaneously hold a key position in Meta's AI sector and serve as a board member of Scale. There are rumors that Alexandr Wang will resign to take on a role as an executive in Meta's newly established "Superintelligence" department, while another insider revealed that discussions have taken place about him becoming Meta's chief AI officer. Meta Sets "Poison Pill" Clause to Prevent Defections, Over $1 Billion in Earnings to be Paid Out Over 5 Years Transaction documents show that the results of Meta's negotiations far exceed what is publicly known, establishing a risk buffer mechanism for its second-largest deal ever. If Scale is sold within the next two and a half years, Meta will have priority to receive full repayment before any other shareholders receive any proceeds. This special shareholder right, known as "liquidation preference," will last until the end of 2027. After receiving substantial dividends funded by Meta, existing Scale shareholders may see their actual returns in the sale fall below expectations, as preferred shares will convert to common stock According to the plan, Meta will invest in Scale and hire Alexandr Wang, but the shares obtained do not include voting rights. However, Meta quickly added new requirements: in the list of terms on May 31, Meta requested to obtain voting rights on most matters except for board elections, priority over other shareholders in profit rights, and the addition of a "poison pill clause," which states that if the company finds Alexandr Wang still participating in Scale's operations during his tenure at Meta, his shareholding ratio will automatically increase. In the recently announced deal, the dividends paid by Meta brought $12.8 billion (approximately RMB 91.9 billion) in unexpected gains to Scale's shareholders. According to insiders, based on his shareholding ratio, Alexandr Wang will receive over $1 billion (approximately RMB 7.2 billion) in profits. Documents show that as long as he continues to work at Meta, most of the cash will be paid out in installments over five years. Zuckerberg's deputies were concerned: Meta invested heavily in a startup that lost its founder, becoming a minority shareholder with limited equity rights. In addition, Scale has recently failed to meet its own financial expectations. Transaction documents and insiders revealed that during behind-the-scenes negotiations, they worked hard to make the terms more favorable to Meta. Meta still needs to obtain approval from U.S. antitrust regulators to participate in company decisions through shares with voting rights. The transaction documents show that Meta agreed to allow Alexandr Wang to exercise its voting rights on most company matters. (A Scale spokesperson denied that the transaction included terms for transferring Meta's voting rights to Alexandr Wang.) A key exception is that Meta requires to retain influence over Scale's sale or financing decisions. If regulatory approval is granted, Meta and Alexandr Wang will jointly control Scale. The structure of this deal reflects Meta's tightrope strategy between finance, strategy, and regulation. Having just undergone significant antitrust litigation, Meta needs to avoid the risk of Washington vetoing the deal while not wanting to lose investment control after paying $14 billion (approximately RMB 100.5 billion). The core driving force behind all this is Zuckerberg's judgment: Meta has fallen behind in the AI race, and Alexandr Wang can turn the situation around. 19 years old entrepreneur, friend of OpenAI founder... Who exactly is Alexandr Wang? Born in 1997 in New Mexico, Alexandr Wang's parents are Chinese immigrant physicists working at the Los Alamos National Laboratory in New Mexico, who deliberately omitted the letter "e" when naming him to make it an 8-character name, reflecting the Chinese cultural significance of "8" as a symbol of wealth. This math prodigy has won numerous awards since childhood, with violin and math competitions forming the main theme of his teenage years. "My arena is mathematics," he summarized in a podcast Alexandr Wang attended Los Alamos High School and graduated a year early, subsequently being admitted to the Massachusetts Institute of Technology (MIT) to pursue a bachelor's degree in mathematics and computer science. However, after entering MIT, he was eager to start a business and ultimately dropped out to move to San Francisco, disregarding his family's academic tradition of having multiple PhDs. "I crave to achieve success faster," Alexandr Wang admitted in a Forbes video. In a TED talk in April 2022, he said, "My parents are distinguished scientists at Los Alamos National Laboratory and have made many contributions in their fields. I want to engage in a career that can match or even surpass that. That's why I decided to become a programmer—I want to change the world." Alexandr Wang's level of social enthusiasm was such that even his former roommate, OpenAI co-founder and CEO Sam Altman, jokingly advised him to tone it down. He is also very close friends with Altman, who, before leading OpenAI, managed the top startup accelerator Y Combinator (YC), and Scale was a member of YC's 2016 cohort. According to mutual friends, the two briefly lived together during the pandemic. A close friend who requested anonymity stated that Alexandr Wang maintains a vast network of contacts by directly reaching out to them, keeping a precise grasp on the work dynamics of contacts at various levels. "There’s no one in the AI circle he doesn’t know," the individual said. Even networking master Altman joked last year about his social intensity: "Certified authority, you (Alexandr Wang) are definitely the person who attends the most parties; it’s practically a full-time job." His network extends far beyond the tech circle. Since founding Scale at the age of 19, Alexandr Wang has formed friendships with the founders of unicorns like Plaid and Figma, while also connecting with Republican lawmakers and right-wing opinion leaders, helping Scale secure numerous government contracts, which may boost Meta's entry into defense technology. A former Biden administration official described him as "authoritative and direct" in Washington meetings and noted that he actively sought to engage with the Biden team. A former employee who accompanied him on visits revealed that Alexandr Wang "almost enjoys" the thrill of being underestimated due to his age. After securing a $90 million contract with the U.S. Department of Defense in 2020 (approximately RMB 6.46 billion), he signed a $250 million (approximately RMB 1.8 billion) federal procurement agreement in 2022, being referred to by foreign media Semafor as "Washington's AI spokesperson." As Scale expands, Alexandr Wang's wealth and prominence have soared in tandem. At 25, Alexandr Wang was named by Forbes as the youngest self-made billionaire with a net worth of $3.6 billion, and after appearing at the Davos Forum, he was also selected as one of Time's 100 Most Influential People of 2024. As the only shareholder of Scale with "1 share, 3 votes" super voting rights, he is set to transition from a paper billionaire to a cash billionaire after receiving cash dividends that vest over five years Alexandr Wang's recent actions mirror those of Mark Zuckerberg, who has also shown goodwill towards Trump. The two even share a similar appearance: both have curly hair and prefer streetwear. The two CEOs have achieved the rare feat of becoming the world's youngest self-made billionaires, and now they are forming a superior-subordinate relationship. Entrepreneurial Inspiration from "Food Being Stolen from the Fridge," Scale AI's Revenue May Reach $2 Billion This Year Before entering MIT, Alexandr Wang took a year off after graduating high school and moved to Silicon Valley, where he found a job as an engineer at the Q&A website Quora. It was at Quora that he met his future co-founder of Scale AI—product head and Thiel Fellow Lucy Guo. ▲Lucy Guo with Alexandr Wang (Source: Forbes) In 2016, 19-year-old Alexandr Wang co-founded Scale AI with Lucy Guo after the summer of his freshman year at MIT, and they were selected for the YC Summer 2016 incubation program (which has incubated companies like Airbnb and Dropbox). He serves as the CEO of Scale AI, while Lucy Guo is responsible for operations and product design teams. YC partner Jessica Livingston recalled in the "Social Radar" podcast that during the interview, Alexandr Wang exhibited a peculiar mix of potential and arrogance. She noted in her notes: "Possibly arrogant, possibly genius—I lean towards the former, but worth investing in." Other partners described this founder using the financial term "high beta," with YC co-founder Paul Graham explaining: "This means either a spectacular failure or conquering the world." They initially submitted an idea for a doctor appointment application, but a few weeks after joining YC, they abandoned the medical application and shifted to machine learning data annotation services. Inspired by AlphaGo's defeat of human chess players, Scale positioned itself as the "Uber of the AI world," recruiting data annotators globally to handle tasks like autonomous driving image recognition. In an interview with Spain's El País, Alexandr Wang revealed that the inspiration for Scale AI came from an idea he had while studying at MIT: he attempted to install a camera in his fridge and develop AI to analyze the footage to find out which roommate was stealing his food. However, due to the large volume of video data, he ultimately failed, but this experiment made him realize the importance of data for the development of AI. Alexandr Wang stated in a TED talk: "To make AI powerful, you need powerful data. And in 2016, when I was still at MIT, high-quality data was extremely scarce." Scale AI has pioneered several key technologies that allow chatbots to behave more human-like and answer complex questions more accurately. Scale AI quickly gained favor from top venture capitalists, but two years after the company's founding, the relationship between the co-founders deteriorated, ultimately leading to Guo Lushi's departure, leaving Alexandr Wang in sole control. Before the release of ChatGPT in 2021, Scale's revenue was only $120 million (approximately RMB 860 million), but it soared to $870 million (approximately RMB 6.2 billion) last year, with the company expecting revenue to reach $2 billion (approximately RMB 14.4 billion) this year. However, Scale's recent performance has fallen short of expectations: it failed to achieve last year's revenue target of $1 billion (approximately RMB 720 million), and the profit-sharing ratio paid to contractors exceeded expectations. Even major clients like Meta have expressed disappointment that Scale did not provide sufficiently high-quality model training data, and competitors like OpenAI have begun to use more suppliers to complete similar work to Scale. Nevertheless, driven by the AI boom, Scale's valuation has still surged. Earlier this year, the company sold employee shares to investors at a valuation of $25 billion (approximately RMB 179.5 billion). Equity is just a side note; Alexandr Wang is the real target of Meta Regarding this deal, a Scale investor revealed that acquiring equity in the startup is "essentially a side condition," and the real goal is Alexandr Wang himself. An executive from a competitor of Meta disclosed that this move stems from Zuckerberg's belief that Alexandr Wang is the "wartime CEO" urgently needed for his "superintelligence" lab. "There are very few companies that can maintain deep relationships with all the top research teams: besides Scale, there's only NVIDIA, and Jensen Huang probably won't accept this job," analyzed Dan Levine, an early investor in Scale and partner at Accel. Zuckerberg hopes that Alexandr Wang's "star effect" can attract top talent from competitors like OpenAI, Anthropic, Google, and Safe Superintelligence (SSI) founded by Ilya Sutskever. According to insiders, Meta is trying to poach SSI co-founder Dan Gross and his investment partner Nat Friedman. Successfully recruiting the two would greatly enhance the strength of Meta's restructured AI department. In the AI race, Alexandr Wang's ability to leverage his network may not be sustainable; his new position will put him directly in competition for talent with figures like Altman and Murati. Altman revealed this week that Meta had offered a signing bonus of $100 million (approximately RMB 72 million) and higher salaries to poach OpenAI developers This massive investment in talent exposes the dilemma of Meta's current AI strategy: its latest model, Llama 4, has performed mediocrely in several performance benchmark tests. The company has yet to determine Alexandr Wang's specific title and his relationship with Meta's head of foundational AI research, Yang Likun. "This is a gamble: the most important thing in the short term is to quickly assemble a research team," a person close to Scale revealed. "They are offering astonishing conditions to poach talent, and Alexandr Wang knows all these top talents... it's like planning a heist." Some contacts believe that Alexandr Wang's social fervor hides opportunism. "Alexandr Wang is a shrewd operator, pursuing maximum fame and fortune," commented a former employee. Another collaborator added, "He is eager to be credited on all projects, even if he hasn't been deeply involved." Scale has refuted such descriptions, emphasizing that Alexandr Wang works diligently and responsibly. Data Annotation Industry Shaken, Scale AI's Competitors Experience Business Surge Scale AI's interim CEO Jason Droege emphasized in an internal memo last Wednesday that this AI startup will not change its established direction due to Meta's massive investment. "Unlike other recent transactions in the AI field, this is neither a business transformation nor a contraction signal," Droege clearly stated in an open letter to clients, employees, and investors. However, some investors still express doubts about the transaction's prospects: "This is bound to create a bizarre future landscape." This deal is good news for Meta and Alexandr Wang, but its benefits for Scale are not obvious, as being tied to Meta may result in the loss of important clients. It is reported that Scale's two major clients and Meta competitors, OpenAI and Google, have begun to reduce collaboration. An OpenAI spokesperson previously stated that over the past 6 to 12 months, the company has determined that Scale cannot meet its increasingly specialized data needs. As OpenAI shifts towards developing advanced AI models that can simulate human reasoning processes and agent models that can perform tasks with limited user input, the requirements for data quality have significantly increased. Forbes previously reported that OpenAI's reduction in collaboration with Scale has been ongoing for months. According to insiders, Microsoft, which is also listed as a client on Scale's official website, currently has no "significant" level of collaboration with Scale. Since these small contracts are not affected by Meta's investment, Microsoft has not yet conducted a business evaluation. Alex Ratner, CEO of competitor Snorkel AI, stated in an interview that after the news of the Meta and Scale AI deal broke, he received inquiries from dozens of potential clients, leading to "tens of millions of dollars" in deals at various negotiation stages. "All responsible LLM (large language model) developers will take action to diversify their suppliers," Ratner said, "The market will see significant changes and new opportunities." Companies including Turing, Invisible Technologies, Labelbox, and Uber Technologies are competing to meet the massive demand for data from AI developers. Many companies claim that their client consultation volume has surged, with some clients concerned that Meta may be privy to their AI research and development processes. "Demand has reached unprecedented levels," said Labelbox founder Manu Sharma, "We have served top AI labs, and now other vendors are also flocking in." "Labs are unwilling to let competitors see their model optimization data," revealed Handshake CEO Garrett Lord, a competitor of Scale, noting that his company's demand "tripled overnight," adding, "Just like Toyota wouldn't allow competitors to tour its production line." Jonathan Siddharth, CEO of another data company Turing, stated that they have added $50 million (approximately 360 million RMB) in potential contracts over the past two weeks, as "leading labs realize that developing AGI (Artificial General Intelligence) requires truly neutral partners." Ryan Kolln, CEO of AI training data company Appen, compared the industry upheaval to "the explosion of the Russia-Europe oil pipeline," with clients urgently seeking alternatives. He added, "After Meta became a major shareholder in Scale, it will be harder to control the ability to obtain intelligence from other foundational model labs." According to insiders, several Scale employees have signed contracts to move to two competing firms in the past week. Conclusion: Will Scale lose its neutrality, and can Meta fulfill its promise of an AI comeback? Scale AI board member Volpi posted on X that "I will miss Alexandr Wang, but Scale will continue to grow," while Accel's Levine released a carefully crafted tribute video. Although the actual impact of Meta's entry on Scale's business remains to be seen, the industry consensus is that this deal will significantly increase market attention on the "invisible segments" of the AI industry chain. Olga Megorskaya, co-founder of data labeling company Toloka AI, pointed out that compared to the highly publicized AI models and computing power chips, data work has long been in a "neglected" state. The potential scale of the financial flows involved in this deal is staggering. According to Garrett Lord, top AI companies invest about $1 billion (approximately 7.2 billion RMB) annually in human data procurement, and the budget is still growing. As competitors scramble to fill the market void left by the Meta deal, this commercial game suggests that the construction of the world's most valuable AI models is undergoing a fundamental reconstruction. Scale AI started with data labeling services, a model that was crucial in the early stages of AI development when companies were still teaching image models to distinguish between cats and dogs or helping language models organize coherent sentences However, with the evolution of AI models, the industry's demand has undergone a dramatic change. The rise of "inference" models has led companies like OpenAI, Google, and Anthropic to primarily seek expert-level data, meaning that the most valuable training data is now written by professionals such as PhDs, who must meticulously document the problem-solving steps for AI to imitate. "The industry is shifting towards requiring increasingly intelligent humans," believes Siddharth, CEO of Turing. "In certain fields, even expert teams need to collaborate to drive progress." The specific task requirements for experts at various AI companies are highly confidential. Industry insiders reveal that while all laboratories will eventually converge on strategies, the longer the confidentiality period, the longer their models can maintain a competitive edge. This also explains why Meta's substantial investment in Scale has made leading AI companies uneasy. Although Meta is currently lagging, if it can acquire the core data secrets of its competitors, it may quickly narrow the gap. Author of this article: Wang Han, Source: ZhDongxi, Original title: "The Secret of a 28-Year-Old Chinese Genius Billionaire: Dropping Out to Start a Business Due to Refrigerator Theft, Wealth Soars with 'AI Labor', Snatched by Silicon Valley Tycoons at High Prices" Risk Warning and Disclaimer The market has risks, and investment should be cautious. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances. Investing based on this is at one's own risk