Meta is putting nearly $15 billion into a massive stake in AI startup Scale AI to boost its AI game and recruit Scale CEO Alexandr Wang.
The plan: Meta will own 49% of Scale AI, but not buy the whole company, to avoid regulator headaches during its FTC antitrust trial. Sources say Wang will lead a new AI research lab at Meta, bringing some Scale colleagues along.
Scale AI started in 2016 and is a key player in prepping data for big AI models at OpenAI, Google, Microsoft — and Meta itself. Valued at around $14 billion, Scale just secured one of San Francisco’s biggest office leases in 2024 and recently partnered with the Department of Defense on "Defense Llama," a customized version of Meta’s open-source Llama model for national security. 
Meta’s AI push has stalled after the lukewarm launch of Llama 4 models. The bigger, more powerful “Behemoth” model is still on hold. Mark Zuckerberg is frustrated competitors like OpenAI are pulling ahead in foundational AI and products.
Meta split its GenAI team after Llama 4’s flop. Connor Hayes now heads AI Products. AGI Foundations is led by Amir Frenkel and Ahmad Al-Dahle, signaling Zuckerberg’s shaken confidence.
Wang, an MIT dropout, is seen as a "wartime CEO" who fits Zuckerberg’s urgency around U.S.-China AI competition.
Alexandr Wang told CNBC in January:
"The United States is going to need a huge amount of computational capacity, a huge amount of infrastructure," Wang said at the time. "We need to unleash U.S. energy to enable this AI boom."
Meta hopes Wang’s deep AI know-how and startup leadership can help it catch up on AI innovation and chatbot design.
A competitor CEO put it bluntly:
"I would say Scale have covered probably 70% of all the models that are built," said Vahan Petrosyan, CEO of SuperAnnotate. With Wang and others from Scale AI, Meta could gain "collective intelligence on how to build a better ChatGPT."
"When Meta is buying them, they’re buying their intelligence," Petrosyan added.
Meta and Scale AI declined to comment.
