Inside the$14.3 Billion Bet on AI Dominance with Talent Poached from OpenAI, Anthropic, and Google
- Mark Zuckerberg has assembled a formidable superintelligence team at Meta, recruiting nearly two dozen top AI researchers from rival firms like OpenAI, Anthropic, and Google.
- With a staggering$14.3 billion investment in Scale AI and the creation of Meta Superintelligence Labs (MSL), Meta is positioning itself as a leader in the AI race under the leadership of Scale AI’s CEO, Alexandr Wang.
- An internal memo obtained by WIRED reveals the names and impressive credentials of the new hires, showcasing Meta’s aggressive push to develop next-generation AI models.
In a bold move that’s shaking up the tech world, Meta CEO Mark Zuckerberg has gone all-in on artificial intelligence, orchestrating a recruiting frenzy that’s pulled in some of the brightest minds from competitors like OpenAI, Anthropic, and Google. Today, Zuckerberg introduced Meta staff to the newly formed superintelligence team via an internal memo, a document obtained by WIRED that lists nearly two dozen researchers with jaw-dropping resumes. This isn’t just a hiring spree; it’s a declaration of intent. Meta is gunning for AI supremacy, and with a$14.3 billion investment in Scale AI, alongside the creation of Meta Superintelligence Labs (MSL), Zuckerberg is betting big on the future.
The memo, first reported by Bloomberg, outlines the structure of MSL, which Zuckerberg describes as encompassing “all of our foundations, product, and FAIR teams, as well as a new lab focused on developing the next-generation of our models.” This ambitious reorganization signals Meta’s shift from being a social media titan to a serious contender in the AI arena. Under the leadership of Alexandr Wang, the CEO of Scale AI who’s now at the helm of MSL, Meta is pooling its resources to innovate at a breakneck pace. While Meta declined to comment on the memo or the hires, the names listed speak volumes about the company’s direction.
Let’s dive into the talent Zuckerberg has brought on board. Trapit Bansal, a pioneer in reinforcement learning on chain of thought and co-creator of the o-series models at OpenAI, is a standout addition. Shuchao Bi, another OpenAI alum, co-created GPT-4o’s voice mode and the o4-mini model, bringing expertise in multimodal post-training. Huiwen Chang, previously at Google Research, invented the MaskIT and Muse text-to-image architectures before co-creating GPT-4o’s image generation capabilities. These are just a few examples of the heavy hitters now wearing Meta’s colors.
The list goes on with names like Ji Lin, who contributed to multiple GPT iterations including 4o, 4.1, and 4.5, as well as the Operator reasoning stack at OpenAI. Joel Pobar, returning to Meta after 11 years of prior service, brings inference expertise from Anthropic and a deep background in performance tooling and machine learning. Jack Rae, formerly of DeepMind, led early LLM efforts like Gopher and Chinchilla, and served as the pre-training tech lead for Gemini. Hongyu Ren, another OpenAI veteran, co-created several models including GPT-4o and o3-mini, while Johan Schalkwyk, a former Google Fellow, contributed to foundational projects like Sesame and Maya.
Rounding out the team are Pei Sun, who worked on coding and reasoning for Gemini at Google DeepMind and built perception models for Waymo; Jiahui Yu, a co-creator of multiple GPT models and a multimodal expert from Gemini; and Shengjia Zhao, who led synthetic data efforts at OpenAI and co-created ChatGPT alongside other landmark models. Notably, the memo doesn’t mention hires from OpenAI’s Zurich office, leaving some speculation about additional talent yet to be revealed.
What does this mean for the tech landscape? Meta’s aggressive poaching and massive financial commitment suggest that Zuckerberg isn’t content to play catch-up in the AI race. By assembling a dream team of researchers who’ve already shaped the industry at places like OpenAI and Google, Meta is positioning itself to challenge the status quo. The creation of MSL, with its focus on next-generation models, hints at innovations that could redefine how we interact with technology, from social platforms to entirely new AI-driven experiences.
From a broader perspective, this move reflects the escalating stakes in the AI arms race. Companies like OpenAI, Anthropic, and Google have long dominated headlines with their breakthroughs, but Meta’s latest play shows it’s ready to disrupt that hierarchy. The$14.3 billion investment in Scale AI isn’t just a financial flex—it’s a strategic alignment with a company known for accelerating AI development through data labeling and infrastructure. With Alexandr Wang steering the ship, Meta could leverage Scale AI’s expertise to scale its ambitions faster than ever.
How will this superintelligence team integrate with Meta’s existing projects? What specific breakthroughs are they targeting with their next-gen models? And perhaps most intriguingly, how will competitors respond to losing such critical talent? For now, the industry watches as Zuckerberg’s vision unfolds, one hire at a time. One thing is clear: Meta isn’t just building a team—it’s building a legacy in AI, and the world is about to see what superintelligence really looks like.