Meta’s landmark investment in Scale AI, a move designed to secure its dominance in the artificial intelligence race, has instead triggered an immediate and severe crisis for its new partner. Google, the largest customer of Scale AI, is now planning to sever ties, and other major tech companies are following suit, according to five sources familiar with the matter who spoke to Reuters.
The exodus is driven by fears that Meta’s $14 billion investment for a 49% stake compromises Scale AI’s neutrality, potentially exposing the proprietary data and strategic roadmaps of Meta’s chief rivals. With Scale AI’s founder, Alexandr Wang, stepping down as CEO to lead a new ‘superintelligence’ lab, competitors are concerned that their research priorities could be laid bare.
This fallout not only threatens Scale AI’s core business but also marks a pivotal moment in the AI arms race, where the independence of critical suppliers has become a paramount concern. What was intended as a strategic coup for Meta has become a seismic shift for the artificial intelligence industry, forcing a realignment of partnerships across Silicon Valley.
An ‘Unbelievable’ Deal
The final terms of the deal, confirmed by Scale AI in an official announcement, underscore the depth of Meta’s strategic urgency. The $14 billion price tag for nearly half the company far exceeded what was reported just days earlier, when initial speculation pointed to a $10 billion investment for a much smaller 10% stake.
The move was so remarkable that Gil Luria, D.A. Davidson’s Head of Technology Research, noted in comments to Yahoo Finance it was an investment “not to even buy a whole company but just to have the head of a company head up your AI effort.” This aggressive acquisition of talent and infrastructure highlights a dramatic pivot for Meta, a company that has historically prioritized building its technology in-house.
Meta’s decision was not made in a vacuum but against a backdrop of significant internal turmoil. The company has been hemorrhaging key personnel, losing 11 of the 14 original authors of its foundational Llama research paper and, in April, its respected head of the Fundamental AI Research (FAIR) group, Joëlle Pineau.
This talent drain has been compounded by significant technical setbacks, including the postponement of its most ambitious model, the 2-trillion parameter Llama 4 “Behemoth,” until at least late 2025.
Having been met with a lukewarm reception when it previously pitched a “Llama Consortium” to competitors for financial help, Meta was forced toward a new strategy: outright acquisition of a key supplier to secure its pipeline. The company is also simultaneously fighting a major copyright lawsuit from authors over the alleged use of pirated books to train its models.
The Price of Lost Neutrality
The consequences of sacrificing neutrality have been swift and brutal. Google, which spent approximately $150 million with Scale in 2024 and had planned to increase that to $200 million this year, has already started conversations with Scale’s rivals. Other major clients, including Microsoft and Elon Musk’s xAI, are also reportedly looking to exit, while OpenAI began pulling back months ago.
Microsoft is now said to be conducting a “comprehensive review” of its relationship with Scale AI, with internal sources expressing concern about conflicts of interest. This has created a massive opening for Scale AI’s competitors, who are seeing a surge in business from AI labs scrambling for independent partners.
The move underscored a new reality in the AI arms race, with Jonathan Siddharth, CEO of rival firm Turing, stating that “The Meta-Scale deal marks a turning point. Leading AI labs are realizing neutrality is no longer optional, it’s essential.” The CEO of another competitor, Handshake, reported that their demand tripled overnight, while the head of Labelbox told Reuters he expects to generate hundreds of millions in new revenue from customers fleeing Scale.
An Alliance Forged in Controversy
The partnership also brings Scale AI’s contentious business practices and deep military ties directly into Meta’s orbit, creating new ethical and regulatory entanglements. As reported by Futurism, Scale AI has been accused of using low-wage overseas labor.
More pointedly, Inc. Magazine has reported on a history of ‘wage theft’ and ‘widespread labor abuses.’ Furthermore, the company is a key contractor for the Pentagon, providing synthetic data environments used to train advanced targeting algorithms for the U.S. military. Scale AI founder Alexandr Wang has been an unapologetic proponent of this military work, framing it as a “‘patriotic duty’ for U.S. tech companies” while asserting there was “‘no room for neutrality in the global technology race.”
The deepening of these ties was underscored by the fact that Meta CTO Andrew Bosworth, was recently sworn in as a lieutenant colonel in a new U.S. Army Reserve innovation unit, alongside other tech executives from Palantir and OpenAI. This fusion of a global consumer data empire with military-grade data machinery could also attract regulatory scrutiny, especially as Meta is already embroiled in a historic antitrust trial with the FTC.
