Another investment with a return of over 1000 times has emerged.

This is likely to become another classic case in Y Combinator's investment history.

According to The Information, Meta, the parent company of Facebook, will acquire a 49% stake in Scale AI for $14.8 billion (approximately RMB 106.29 billion). This deal will be Meta's second-largest transaction since its $19 billion acquisition of WhatsApp.

Scale AI is a data annotation company. Google, Microsoft, OpenAI, xAI and other direct competitors of Meta are all major clients of Scale AI. The company's valuation after a new round of financing last year was approximately $13.8 billion. After Meta's investment, its valuation soared to $29 billion, doubling directly.

Recently, the technology merger and acquisition market in the United States has been heating up. Some star AI projects have received high-premium acquisition offers from buyers, contributing many exit cases with a hundred-fold return for early investors behind these projects.

The story of Scale AI continues this classic plot. As early as its establishment in 2016, Scale AI received a $120,000 seed-round financing from Y Combinator. Although Y Combinator's shareholding ratio has significantly decreased after multiple rounds of financing dilution, this early investment has basically locked in a return of over 1000 times for it.

Chinese-American Genius Drops Out of School to Start a Business, with a Valuation of $29 Billion

Chinese-American genius, dropping out of a prestigious university, the world's youngest billionaire... Any one of these labels is enough to make Alexandr Wang, the founder of Scale AI, a focus of attention.

Alexandr Wang, 28 years old this year, grew up in a Chinese-American family in Los Alamos, New Mexico. His parents are physicists at the local national laboratory. Influenced by the strong scientific atmosphere at home, Alexandr Wang showed extraordinary talent in mathematics and programming from a young age. He participated in the national mathematics Olympiad and physics competitions many times during his school years.

It seems that every successful entrepreneur in Silicon Valley has an unorthodox educational experience, and Alexandr Wang is no exception.

At the age of 17, Alexandr Wang graduated from high school early and received an admission letter from the Massachusetts Institute of Technology (MIT). However, instead of going to the East immediately, he went to Silicon Valley and worked as a software engineer at the Q&A SNS website Quora. It was this internship experience that allowed him to meet Lucy Guo, the later co-founder of Scale AI. (Regarding Lucy Guo, China Venture Capital has introduced her in detail in the article "She Surpasses Taylor Swift to Become the World's Youngest Female Billionaire".)

After that, Alexandr Wang briefly returned to the educational track and entered MIT to study mathematics and computer science. He later publicly recalled, "Although there are hundreds of talented and gifted students at MIT, no one has successfully built anything with AI. We were all researching artificial intelligence, but we all encountered a major bottleneck - there was no good data."

This perception deeply influenced him. In 2016, Alexandr Wang decided to drop out of college and co-founded Scale AI with Lucy Guo, choosing the business direction of AI training data annotation.

In both past and present contexts, data annotation has never been a glamorous business. Compared with the highly - regarded large models, data annotation has always been regarded as "dirty and tiring work" and an area that AI giants are reluctant to get involved in. However, as the "fuel" for training deep - learning algorithms, data annotation is indispensable for the entire AI industry.

Peter Thiel, a Silicon Valley tycoon, once highly praised Scale AI: "In the fierce competition, AI companies will appear and disappear, but Scale will always be there. Because Scale AI provides the infrastructure for the entire AI/ML (artificial intelligence/machine learning) industry, and data is the most important thing in this industry."

Scale AI's business started in the field of autonomous driving, mainly providing AI data annotation outsourcing services for companies such as Cruise and Tesla. Later, the company's business map continued to expand and it gradually transformed into an enterprise - level data integrator, assisting customers in AI development and deployment.

Today, many technology and manufacturing giants such as OpenAI, Google, Microsoft, Etsy, General Motors, PayPal, Samsung, Toyota, and Uber are all important clients of Scale AI. The company also cooperates with public sectors such as the US Department of Defense and the Qatari government, participating in military and social projects.

According to public data, in 2023, Scale AI achieved an annualized revenue of $750 million, a three - fold increase year - on - year, and became one of the highest - revenue AI startups at that time.

Alexandr Wang's outstanding performance has also made him well - known in the US AI circle. In September 2023, Time magazine released its first list of the world's top 100 AI figures, and he was named a global AI leader. In April 2025, with a net worth of $2 billion, he was included in the list of "The World's Youngest Billionaires in 2025" selected by Forbes China.

If Meta's acquisition can be successfully completed, the net worth of this 28 - year - old "Chinese - American genius" will also increase significantly.

Meta's "AI Anxiety"

Meta's decision to spend $14.8 billion to acquire Scale AI is interpreted by the outside world as a high - stakes gamble to relieve its "AI anxiety".

Back in February 2023, Meta released a large language model (LLM) called LLaMA. This move was regarded as a groundbreaking event for open - source large language models at that time and established Meta's leading position in the field of AI large models.

However, with the all - around encirclement of competitors such as OpenAI, Google, and DeepSeek, Meta's first - mover advantage is gradually disappearing. A landmark event occurred in April this year when Meta launched the open - source next - generation artificial intelligence model Llama 4. The product's release did not receive the expected popularity. Instead, it performed far below expectations. Not only did it lag behind its competitors in many key indicators, but it was also questioned for "over - promising and under - delivering".

In fact, even Zuckerberg himself has realized that Meta is gradually falling behind in the AI arms race. According to foreign media reports citing people familiar with the matter, Zuckerberg is dissatisfied with the company's development status in the field of artificial intelligence and is personally recruiting about 50 AI experts to form a super - intelligent team dedicated to achieving artificial general intelligence (AGI).

This investment in Scale AI is a key layout for Zuckerberg to build this team. In a public all - hands letter, Alexandr Wang, the founder of Scale AI, mentioned that he would step down as the CEO of Scale AI and join Meta with some employees. It is reported that Alexandr Wang will join the super - intelligent team personally built by Zuckerberg and hold an important position.

It is worth mentioning that the significance of this acquisition is not only the introduction of talent. The integration of Scale AI will also help Meta improve its key technologies. According to industry analysis, through this acquisition, Meta is expected to reduce the training data pollution rate from 15% to 2% and shorten the training cycle of the next - generation Llama 5 model by 40%.

It can be seen that this deal is crucial for Meta at present.

Seed Investors Get a Return of Over 1000 Times

Of course, Alexandr Wang and Meta are not the only ones benefiting from this huge transaction. Multiple venture capital institutions behind Scale AI will also receive substantial returns.

According to CVSource data, as a hot AI unicorn in Silicon Valley, Scale AI has completed a total of seven rounds of financing since its establishment, six of which occurred before 2022.

As early as the company's establishment in 2016, Y Combinator provided a $120,000 seed - round investment to Scale AI and continued to increase its investment in the A, B, E, and F rounds. As Scale AI's valuation reaches $29 billion, Y Combinator's earliest seed - round investment has basically locked in a book return of over 1000 times.

In addition, Accel, an established venture capital firm in Silicon Valley, is also one of the early investors in Scale AI. In 2017, Accel led Scale AI's A - round financing, and Daniel Levine, a partner of the institution, joined Scale's board of directors. In the subsequent B, C, and F rounds of financing, Accel also appeared on the list of investors. It is reported that after the completion of this acquisition, Accel will receive an exit benefit of over $2.5 billion (approximately RMB 18 billion).

The investment experience of Paige Craig, the managing partner of Outlander VC, another shareholder of the company, is even more legendary. According to his public disclosure, when Scale was still in its early startup stage, he accidentally learned about the entrepreneurial story of co - founder Lucy Guo from an Uber driver. Later, he contacted Lucy Guo and finally decided to invest approximately $245,000 when the company's valuation was only $3 million. Based on this estimate, Paige Craig will also get a return of several hundred times on his investment.

In addition to the above institutions, investors in Scale AI also include Index Ventures, Tiger Global Management, the Founders Fund managed by Silicon Valley tycoon Peter Thiel, etc.

Scale AI's latest round of financing took place in May 2024. At that time, Scale AI announced that it had received $1 billion in financing from as many as 20 investors. In addition to first - tier investment institutions and old shareholders, there were also super - large companies such as Amazon, Nvidia, Intel, Meta, and AMD.

It is reported that Scale AI's valuation at that time was approximately $13.8 billion (approximately RMB 100.2 billion). This means that even late - stage investors are expected to get their principal back and obtain excess returns.

This article is from the WeChat official account “Dongsi Shitiao Capital” (ID: DsstCapital). The author is Wang Manhua, and it is published by 36Kr with authorization.