Meta is reportedly investing nearly $15 billion in Scale AI, acquiring a 49% stake in the data-labeling firm and bringing CEO Alexandr Wang onboard to co-lead a new “superintelligence” lab. The move represents one of Meta’s largest AI investments to date, signaling a strategic pivot to boost its AI innovation capabilities and compete with leaders like OpenAI, Google DeepMind, and Anthropic.
The partnership underscores a critical shift: Meta is betting not on social platforms but on high-quality training data — the foundational fuel for frontier AI models. Scale AI, a long-standing provider of annotated data for major AI labs, has recently focused on hiring PhDs and senior engineers to generate increasingly complex, high-quality datasets.
Key takeaways:
- Meta’s Llama 4 model underperformed in comparison to peers, prompting internal concerns over lagging innovation.
- The company also faces brain drain, losing 4.3% of its top AI talent in 2024, per SignalFire.
- Wang’s leadership, though lacking deep AI research credentials, is seen as a bold move to reinvigorate Meta’s generative AI ambitions.
While the partnership could offer Meta exclusive access to next-gen training data, it also introduces market risks. AI labs that previously relied on Scale may turn to neutral data providers like Turing and Surge AI amid concerns over conflicts of interest. Meanwhile, the rise of synthetic data and in-house annotation efforts may further complicate Scale AI’s post-deal positioning.
As OpenAI prepares to release GPT-5 and other competitors accelerate, Meta’s ambitious move reflects a high-stakes attempt to reclaim AI relevance. However, whether this bet will deliver long-term advantage or alienate existing partners remains an open question in the fast-evolving AI landscape.
Source:
https://techcrunch.com/2025/06/11/can-scale-ai-and-alexandr-wang-reignite-metas-ai-efforts/
