Meta’s $10B Scale AI Deal and New Superintelligence Team Fuel the AGI Race

Yesterday, billionaire and Meta Platforms CEO Mark Zuckerberg announced the formation of a new expert team called Superintelligence to pursue artificial general intelligence (AGI), a machine capable of matching or surpassing human abilities.

Meta has invested more than $10 billion in Scale AI, and the founder Alexandr Wang is expected to join the AGI group, marking the largest external AI investment in Meta’s history and highlighting a shift toward high‑quality data pipelines rather than merely complex algorithms.

The investment also deepens both companies’ collaboration with the U.S. government, as they co‑develop Defense Llama, a military‑grade version of Meta’s Llama model, and Scale AI has secured a contract with the Department of Defense to create AI agents for combat.

This government partnership adds strategic value beyond direct financial returns, providing stable, long‑term revenue and positioning the firms as essential infrastructure providers for national AI capabilities.

Zuckerberg, now in “founder mode,” plans to recruit about 50 specialists for the Superintelligence team, including a new AI research lead to accelerate AGI development.

The move follows disappointment with the performance and reception of Meta’s open‑source Llama 4 model, prompting a delay of the flagship large language model Behemoth despite claims of superiority over competitors.

Competitors such as OpenAI are also seeking large investments to advance AGI, intensifying competition across the industry.

Meta is allocating hundreds of billions of dollars this year to AI projects, with Zuckerberg promising “trillions of dollars” in investment over the coming years, citing the company’s robust advertising revenue to fund massive data‑center construction for world‑leading AI compute.

By 2025, Meta’s stock has risen 19 %, the largest gain among the seven major tech stocks so far this year.