In 2025, Meta CEO Mark Zuckerberg has intensified his efforts to bolster Meta’s artificial intelligence (AI) capabilities by aggressively recruiting top researchers from rival companies, including OpenAI and Google DeepMind. Offering massive compensation packages, some exceeding $100 million, Zuckerberg is personally leading a talent acquisition drive to build a new “superintelligence” team aimed at catching up in the AI race. While his strategy has secured some high-profile hires, it has also sparked criticism and highlighted Meta’s challenges in AI innovation. Below is an overview of Zuckerberg’s poaching efforts and their implications.
Zuckerberg’s campaign began with a hands-on approach, reportedly contacting hundreds of AI experts via WhatsApp and hosting private dinners at his homes in Palo Alto and Lake Tahoe. His targets included Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai, three key researchers who helped establish OpenAI’s Zurich office. These researchers, previously with Google DeepMind, joined Meta’s superintelligence team in June 2025, marking a significant win. Meta also recruited Trapit Bansal from OpenAI and Jack Rae from DeepMind, along with Scale AI’s CEO Alexandr Wang as part of a $14.3 billion deal. These hires aim to strengthen Meta’s pursuit of advanced AI systems.
The scale of Zuckerberg’s offers is staggering. OpenAI CEO Sam Altman revealed that Meta dangled signing bonuses as high as $100 million, with annual compensation packages reaching seven to nine figures. Zuckerberg’s “Recruiting Party” WhatsApp group, involving senior Meta executives, coordinates these efforts, targeting top talent through research papers and personal outreach. However, not all attempts succeeded. OpenAI co-founders Ilya Sutskever and John Schulman, along with Google’s Koray Kavukcuoglu, rejected Meta’s offers, and Altman claimed none of OpenAI’s “best people” left, suggesting loyalty to mission driven companies outweighs money for some.
Zuckerberg’s poaching spree is driven by Meta’s struggles to keep pace with AI leaders like OpenAI, Google, and DeepSeek. Meta’s AI models, like Llama 4 and Behemoth, have underperformed, and its Fundamental AI Research (FAIR) team lost key members, including leader Joelle Pineau in April 2025. Facing pressure from a declining ad business and competition from TikTok, Meta is projected to spend $65 billion on AI infrastructure in 2025. Zuckerberg’s strategy described as “copy, poach, repeat” by critics also includes failed acquisition attempts of startups like Safe Superintelligence ($32 billion) and Perplexity AI, showing his willingness to buy talent and technology.
Critics, including Altman, argue that Zuckerberg’s approach undermines innovation. Altman, on the Uncapped podcast, called the offers “crazy” and said heavy spending can’t replicate a culture of creativity, predicting Meta’s focus on poaching will lead to a weaker team dynamic. Some researchers also distrust Meta after its January 2025 layoffs of 3,600 employees, which raised concerns about job security. Despite these challenges, Zuckerberg’s defenders note that Meta’s open-source Llama models have fueled third-party innovation, and his aggressive hiring signals a bold pivot toward advanced AI, like reasoning agents for business and consumer apps.
The poaching war has broader implications. The competition for a limited pool of elite AI researchers estimated at fewer than 1,000 globally is driving up salaries and accelerating innovation across healthcare, finance, and more. However, it risks concentrating talent in a few tech giants, potentially stifling smaller players. For Meta, success depends on whether Zuckerberg’s high stakes bets translate into groundbreaking AI. For now, his recruiting blitz shows both ambition and desperation in the race for superintelligence.
Key Points of Zuckerberg’s Poaching Efforts
1. High-Profile Hires: Meta poached Lucas Beyer, Alexander Kolesnikov, Xiaohua Zhai, and Trapit Bansal from OpenAI, plus Jack Rae from DeepMind and Scale AI’s Alexandr Wang.
2. Massive Offers: Zuckerberg offered $100 million signing bonuses and seven- to nine-figure annual packages, targeting top researchers via WhatsApp and dinners.
3. Mixed Success: While some joined Meta, OpenAI’s Ilya Sutskever, John Schulman, and Google’s Koray Kavukcuoglu declined, prioritizing mission-driven work.
4. Meta’s AI Struggles: Underperforming models like Llama 4 and Behemoth, plus FAIR team losses, drive Zuckerberg’s urgent talent grab.
5. Criticism and Impact: Altman criticizes Meta’s “copy and poach” strategy for lacking innovation, but the talent war is speeding up AI advancements industry-wide.