Meta Platforms Chief Executive Officer Mark Zuckerberg has rolled the company’s artificial-intelligence operations into a single division dubbed Meta Superintelligence Labs and declared that building AI systems that exceed human capabilities is now the firm’s primary engineering goal. An internal memo dated 30 June placed all research, product and infrastructure teams under the new umbrella. The unit will be headed by Alexandr Wang, formerly of Scale AI, and ex-GitHub chief Nat Friedman. Meta simultaneously recruited 11 senior researchers—several from OpenAI and DeepMind—as part of what Zuckerberg called an effort to assemble “the most elite and talent-dense team in the industry.” On 14 July, the billionaire founder said Meta will spend "hundreds of billions of dollars" to build a series of multi-gigawatt data-centre superclusters. The first complex, Prometheus, is scheduled to come online in 2026, while a second, Hyperion, could scale to five gigawatts. The company has raised its 2025 capital-expenditure forecast to between $64 billion and $72 billion and has already invested $14.3 billion for a 49% stake in Scale AI to secure additional expertise and infrastructure. In a 30 July open letter titled “Personal Superintelligence,” Zuckerberg said Meta’s models have begun to display early self-improvement and that delivering personal superintelligence to billions of users—likely through devices such as AI-enabled smart glasses—is within reach. He cautioned that new safety risks may limit how much of the technology the company decides to open-source. Meta generated almost $165 billion in revenue last year and says its advertising cash flow will fund the AI push without outside capital. With Meta AI already serving more than one billion monthly users, the escalating investment places the social-media giant in direct competition with OpenAI and Alphabet’s Google in the race to dominate next-generation artificial intelligence.