Scale AI, a startup specializing in providing high-quality data for training A.I. systems, is making inroads into the Big Tech-dominated A.I. race. Bloomberg reported on Sunday (June 8) that Meta is in talks to invest more than $10 billion in the San Francisco-based company, a deal that could be one of the largest private company financing rounds ever. If finalized, it would also significantly increase the wealth of Scale AI’s young founders, particularly CEO Alexandr Wang.
As part of the deal, which hasn’t been officially announced, Meta is reportedly recruiting Wang to lead a new internal research lab. This team will focus on advanced A.I. technologies, such as artificial general intelligence (A.G.I.), with plans to integrate A.G.I. into products like Meta A.I. and Ray-Ban Meta A.I. Glasses. The lab is expected to hire around 50 experts in the field.
Meta and Scale AI did not respond to requests for comment from Observer.
Meta’s collaboration with Scale AI coincides with the company’s plan to double down on A.I. in 2025, entering a full-throttle race for A.G.I. with other Silicon Valley giants. Meta plans to invest up to $72 billion in A.I. over the next year, significantly expanding its chip stockpile, data center capacity and A.I. product offerings. “This is going to be intense,” Zuckerberg said in January, describing 2025 as a pivotal year to “set the course for the future.”
Real Also: Meet Lucy Guo, the New Youngest Self-Made Woman Billionaire Dethroning Taylor Swift
In recent months, Wang has become an increasingly prominent figure in Washington as the U.S. strives to maintain its lead in A.I. Beyond warning about the rise of Chinese A.I. startup DeepSeek, he appealed directly to the Trump administration for A.I. investments and, in April, testified before Congress, urging the government to establish a national A.I. data reserve.
Wang’s focus on federal strategies is rooted in his upbringing in Los Alamos, N.M., where he was raised by two weapons physicists. After stints at tech ventures Addepar and Quora, Wang—much like Mark Zuckerberg—dropped out of university to start his own company. In 2016, he co-founded Scale AI with Lucy Guo, another former Quora staffer who left the company in 2018.
What to know about Scale AI and Alexandr Wang
Scale AI recruits large teams of human contractors worldwide to label text, image and video data for fine-tuning A.I. systems. The company initially focused on autonomous vehicles before pivoting to generative models. With contributors in around 9,000 locations globally, Scale AI has faced labor practice criticism, including allegations of wage theft and worker misclassification. Earlier this year, the U.S. Department of Labor closed a nearly year-long investigation into its employee practices.
As A.I. companies struggle to find high-quality data sources for their increasingly sophisticated models, Scale AI has shifted focus to expert contributors—such as doctors, lawyers and Ph.D.-holders—to provide more specialized data. Its clients include major players like OpenAI, Uber, and government agencies such as the U.S. Department of Defense (DoD). Scale AI’s work with the DoD includes developing an evaluation framework for A.I. systems and integrating A.I. agents into military and operational planning.
Beyond Washington, Wang maintains close relationships with influential figures across Silicon Valley. He has been friends with OpenAI CEO Sam Altman for years, having roomed with him during the Covid-19 pandemic, and launched Scale AI through Y Combinator, the startup accelerator once led by Altman.
Both OpenAI and Meta, along with other investors like Nvidia and Amazon, backed Scale AI during a 2024 funding round that valued the startup at nearly $14 billion. Scale AI’s 2024 revenue was reported at $870 million and is projected to more than double to $2 billion this year.
A partnership with Scale AI would position Meta alongside Big Tech rivals that have made similar moves. For instance, Microsoft invested $13 billion in OpenAI, while Amazon and Google have poured billions into A.I. startup Anthropic.