In a stunning escalation of the global AI talent war, Meta has successfully poached three key researchers from OpenAI’s Zurich office, a direct recruiting coup against its chief rival. The move, confirmed in a report by The Wall Street Journal, lands a significant victory for CEO Mark Zuckerberg as he aggressively attempts to steer Meta out of a period of deep internal turmoil and strategic setbacks in its race for artificial general intelligence.
The researchers—Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai—were instrumental in establishing OpenAI’s Swiss presence late last year and previously worked together at Google’s DeepMind. Their departure directly contradicts recent confident statements from OpenAI CEO Sam Altman. Just last week, Altman claimed that while Meta was making aggressive offers with nine-figure signing bonuses, his “best people” had not left. In a public appearance Tuesday, even as the hires were being finalized, Altman struck a defiant tone, stating, “It’s like OK, Zuckerberg is doing some new insane thing. What’s next?”
This high-profile poach is the latest and most tangible result of an audacious “buy or poach” strategy Meta has deployed to fix its flagging AI efforts. The move highlights the immense pressure on Zuckerberg to acquire top-tier talent by any means necessary, a symptom of a company grappling with significant internal development hurdles and intense competitive pressure. It demonstrates a willingness to spend billions and upend industry alliances to plug innovation gaps.
The ‘Buy or Poach’ Playbook
Meta’s aggressive hiring spree appears to be a direct consequence of being repeatedly rejected on the acquisition trail. The company’s playbook became clear after its attempt to acquire generative video startup Runway was rejected. That rejection prompted a strategic pivot toward the controversial “acqui-hire” of Scale AI’s founder.
This was not an isolated incident. Meta has reportedly held unsuccessful takeover discussions with other key players, including AI-native search engine Perplexity and Ilya Sutskever’s $32 billion startup, Safe Superintelligence (SSI). Zuckerberg had also tried to take over ex-OpenAI CTO Mira Murati’s Thinking Machines Lab. Unable to buy the companies it coveted, Meta shifted to hiring their leadership. Following the failed SSI bid, the company entered talks to hire prominent AI investors Nat Friedman and Daniel Gross.
This pattern reveals a company that, when unable to purchase an asset, is willing to spend vast sums to acquire its most valuable component: its people. The strategy extends beyond just poaching from OpenAI, with AInvest reporting that Meta has previously hired other top talent from labs like DeepMind.
This desperate hunt for talent is underscored by hard data. According to SignalFire’s 2025 State of Tech Talent Report, Meta’s two-year AI talent retention rate sits at 64%, lagging behind rival Anthropic’s 80% and roughly on par with OpenAI’s 63%. The successful hires have reportedly sent ripples through its rival, with one anonymous OpenAI researcher saying, “There’s genuine shock.”
Trading One Crisis for Another: The Scale AI Gambit
Nowhere is Meta’s high-risk strategy more evident than in its partnership with Scale AI. The company finalized a colossal investment of $14 billion for a 49% stake in the data-labeling firm, primarily to install its founder, Alexandr Wang, as the head of Meta’s new superintelligence lab. It was an investment made not to acquire an entire company, but simply to bring in the CEO of a company to lead inhouse AI initiatives. However, the move immediately backfired, igniting a crisis of confidence among Scale AI’s other Big Tech clients who feared its neutrality was compromised.
The fallout was swift, with reports that Google, Scale’s largest customer, began planning to sever a contract worth hundreds of millions. The exodus forced Scale AI’s new interim CEO to issue a public letter to customers and employees, insisting the company remains independent. The episode underscored a new industry reality, with Turing CEO Jonathan Siddharth stating that for leading AI labs, “neutrality is no longer optional, it’s essential.”
Compounding the crisis, recent reports expose major security lapses at Scale AI that were far more severe than initially understood. Scale AI exposed sensitive personal information about its own contractors like Google and xAI in spreadsheets titled “Good and Bad Folks,” with labels like “low quality” and “cheating.” The leak exposed confidential client projects, including details on how Google uses OpenAI’s ChatGPT to fine-tune its own models and information about Elon Musk’s plans for xAI. The revelations of such fundamental security flaws now raise serious questions about the oversight and due diligence Meta performed before its multi-billion-dollar commitment.
Internal Chaos Fuels External Aggression
Meta’s audacious external moves are a direct response to a firestorm of internal challenges. The company has been hemorrhaging the talent behind its foundational AI work, losing 11 of the 14 original authors of its Llama research paper. These personnel issues have been compounded by major technical setbacks, including the significant postponement of its most ambitious model, Llama 4 “Behemoth,” after it underperformed on key benchmarks.
This pressure has fueled what anonymous Meta engineers on the platform Blind described as a “panic mode” inside the company, with frantic efforts to reverse-engineer competitor models. According to one engineer, “Management is worried about justifying the massive cost of GenAI org.” In a physical demonstration of priority, Mark Zuckerberg has reportedly reorganized Meta’s headquarters to position the new superintelligence team directly adjacent to his own office, which is basicallz a spatial declaration of priority for the initiative. This all-in approach is a hallmark of Zuckerberg’s leadership style.
This internal chaos provides the crucial context for Meta’s current strategy. The poaching of the OpenAI trio is not a random act of corporate rivalry but the most direct and successful manifestation of a desperate, high-stakes campaign. While Meta has secured a critical short-term talent victory, its chaotic approach of spending lavishly to trade one crisis for another raises serious questions about its ability to build a stable, long-term foundation for AI leadership. The company may have acquired the architects it wanted, but it has yet to prove it can build a house on solid ground.
