Editor’s take: Big tech companies are locked in a fierce race to develop artificial intelligence that matches or exceeds human intelligence – a breakthrough with the power to transform industries and society. Meta’s new partnership with Scale AI underscores just how urgent and high-stakes this competition has become, as the company commits to unprecedented investments to keep pace.
Insider sources told The Information that Meta is making a bold move to reshape its artificial intelligence ambitions by investing $14.8 billion for a 49 percent stake in Scale AI, a San Francisco – based data labeling firm. Although the companies have yet to finalize the deal, it ranks among the largest external investments in Meta’s history and signals a dramatic shift in the company’s strategy as it aims to regain ground in the rapidly evolving AI sector.
Meta structured the deal so its cash goes to Scale AI’s existing shareholders, including principal venture capital backers and employees. Meanwhile, Scale AI CEO Alexandr Wang will take on a senior executive role at Meta, leading a newly formed research lab focused on “superintelligence” – an advanced form of artificial intelligence that aims to surpass human cognitive abilities. Several Scale AI employees will join Meta as part of this initiative, designed to accelerate the company’s AI development after a series of recent setbacks.
Meta CEO Mark Zuckerberg has reportedly taken a hands-on approach to the company’s AI pivot, personally reaching out to researchers and spearheading recruitment of the new superintelligence team. The company plans to hire around 50 experts for the group, reflecting Zuckerberg’s determination to reposition Meta at the forefront of artificial general intelligence. The effort follows the underwhelming rollout of Meta’s recent AI models, including the latest version of its Llama language model, which faced delays and fell short of expectations – frustrating senior leadership and sparking internal talks of management shakeups.
Unlike other artificial intelligence startups developing large language models, Scale AI focuses on delivering the massive, high-quality labeled datasets essential for training advanced systems. The company operates a global network of more than 100,000 contractors who annotate images, write text, and generate training material for clients that include OpenAI, Google, and Meta. In recent years, Scale AI has expanded into software tools that help businesses build AI solutions, but data labeling remains its core business.
Meta’s investment in Scale AI reflects a broader trend among tech giants increasingly acquiring minority stakes in promising startups to secure talent and technology while avoiding heightened antitrust scrutiny. Microsoft, Amazon, and Google made similar moves last year, but Meta’s partnership with Scale AI stands out for its focus on data infrastructure rather than model development.
People familiar with the deal told The Wall Street Journal that Meta would receive nonvoting shares in Scale AI, with a provision allowing those shares to convert into voting stock – potentially giving Alexandr Wang control. Part of the investment will go directly to existing shareholders, who will retain their current stakes in the company.
