Scale AI’s Alexandr Wang Poised to Recharge Meta’s AI Ambitions

Meta is reportedly making a nearly $15 billion investment in the data-labeling company Scale AI, acquiring a 49% stake and bringing on CEO Alexandr Wang to help lead a new internal "superintelligence" lab.

This massive deal echoes Meta’s history of bold, high-stakes investments, like its $19 billion WhatsApp acquisition and $1 billion Instagram purchase. Each of those deals was initially criticized as an overpayment, and the reaction to the Scale AI news has followed a similar pattern, leaving many investors and founders puzzled.

Ultimately, WhatsApp and Instagram became cornerstones of Mark Zuckerberg’s empire. The critical question now is whether the Scale AI investment will prove equally strategic, showcasing Zuckerberg’s foresight once more, or if it represents a desperate attempt to keep pace with AI leaders like OpenAI, Google, and Anthropic.

This time, Meta’s bet isn't on a trendy social app, but on the foundational data used to train cutting-edge AI models. For years, top AI labs, including OpenAI, have depended on Scale AI for data production and annotation. Recently, Scale AI and its rivals have begun hiring highly qualified personnel, such as PhD scientists and senior engineers, to create premium data for frontier AI research.

A close partnership with a data provider like Scale could be advantageous for Meta. Sources indicate that Meta’s leadership has expressed concerns about a lack of data innovation within its core AI teams.

Earlier this year, Meta’s generative AI unit released the Llama 4 model family, which failed to match the performance of Chinese lab DeepSeek’s models and was widely seen as underwhelming. Compounding these challenges, Meta is facing a brain drain, having lost 4.3% of its top talent to rival AI labs in 2024, according to SignalFire data.

Meta’s strategy to revitalize its AI efforts hinges not only on Scale AI but also on CEO Alexandr Wang leading the new superintelligence team. The 28-year-old founder has built a strong reputation in Silicon Valley as an ambitious, well-connected, and persuasive leader. Recently, he has engaged with global leaders to discuss AI's societal implications.

Save $200+ on your TechCrunch All Stage pass

Build smarter. Scale faster. Connect deeper. Join visionaries from Precursor Ventures, NEA, Index Ventures, Underscore VC, and beyond for a day packed with strategies, workshops, and meaningful connections.

Save $200+ on your TechCrunch All Stage pass

Build smarter. Scale faster. Connect deeper. Join visionaries from Precursor Ventures, NEA, Index Ventures, Underscore VC, and beyond for a day packed with strategies, workshops, and meaningful connections.

However, Wang lacks prior experience leading an AI research lab and does not share the deep research background of peers like Safe Superintelligence’s Ilya Sutskever or Mistral’s Arthur Mensch. This may explain why Meta is also reportedly recruiting established AI researchers like DeepMind’s Jack Rae for the new group.

Scale AI’s long-term role post-investment is uncertain. The importance of real-world data for AI training is evolving, with some labs bringing data work in-house and others increasingly using synthetic, AI-generated data. In April, The Information reported that Scale AI had missed some revenue targets.

According to Anyscale co-founder Robert Nishihara, several leading AI labs are pioneering new, compute-intensive methods for leveraging and optimizing data.

“Data is a moving target,” Nishihara told TechCrunch. “It’s not just a finite effort to catch up — you have to innovate.”

Meta’s deepened involvement with Scale AI may deter other AI labs that have been its traditional clients. If so, the deal could benefit Scale AI’s competitors, such as Turing, Surge AI, and newer entrants like the recently launched LM Arena.

Turing CEO Jonathan Siddharth told TechCrunch via email that customer interest has risen following rumors of the Meta-Scale AI deal.

“I think there will be some clients who will prefer to work with a partner that’s more neutral,” he said.

Only time will reveal the outcome of Meta’s massive investment. The company has considerable ground to regain in the AI race, and its competitors are not pausing. OpenAI is preparing to launch its next flagship model, GPT-5, along with its first openly available model in years—a direct challenger to Meta's Llama lineup.