Meta has reportedly made a striking move in AI by investing close to $15 billion in data-labeling firm Scale AI, acquiring a 49% stake, and hiring its founder and CEO Alexandr Wang to lead a new internal “superintelligence” lab.
This substantial investment has sparked comparison to previous high-profile deals by Meta, including its $19 billion acquisition of WhatsApp and its $1 billion purchase of Instagram. Both deals initially faced criticism for their steep valuations but eventually became pivotal to Meta’s success. Observers are now debating whether the Scale AI partnership represents a similarly visionary strategy or if it signals Meta’s desperation to catch up with competitors such as OpenAI, Google, and Anthropic.
In contrast to Meta’s prior acquisitions of popular social media platforms, this time the company is betting big on data—the critical component behind state-of-the-art AI models. Scale AI has been a major supplier of training data and labeling services for top-tier AI research labs, including OpenAI. Recently, it has strengthened its position by hiring distinguished PhD and senior engineering talent to further improve the quality of the datasets.
Meta hopes to leverage Scale AI’s expertise to address internal shortcomings around data innovation. Notably, Meta’s latest generative AI model launch, Llama 4, failed to meet expectations, trailing significantly behind models from rival labs like China’s DeepSeek. Internal morale and retention have also been challenges, with Meta losing approximately 4.3% of its top talent to competing AI labs last year alone.
Beyond Scale AI’s data capabilities, Meta is banking on Wang himself to reignite momentum. At 28, Wang has distinguished himself as an ambitious and connected entrepreneur, frequently consulting with global leaders about AI’s societal impacts. Yet Wang’s appointment to lead an elite AI research lab represents new territory for the young executive, who does not currently boast the traditional high-level AI research experience of counterparts at labs such as Safe Superintelligence or Mistral.
To bolster Wang’s team, Meta reportedly aims to recruit high-profile researchers, including DeepMind’s Jack Rae. The exact fate of Scale AI after integration into the Meta structure remains uncertain. The changing nature of AI data collection—where labs increasingly rely on synthetic outputs rather than purely human-annotated real data—adds complexity to the future usefulness of Scale’s business model, particularly as the firm recently missed revenue projections.
Some industry insiders speculate that Scale AI’s close tie-up with Meta could inadvertently benefit its competitors. Firms like Turing and Surge AI, and newcomers such as LM Arena, may position themselves as more attractive partners for customers wary of potential conflict of interests or data confidentiality concerns.
Ultimately, the massive investment in Scale AI reflects high stakes for Meta. With intense competition from OpenAI—set to introduce the highly anticipated GPT-5—Meta’s ability to regain a leading position in AI research and product innovation remains uncertain. Whether the Scale AI acquisition will become another prescient Mark Zuckerberg-led strategic maneuver, or an expensive misstep, is a narrative yet to unfold.
