Meta launches massive AI infrastructure push under new compute strategy

Meta Platforms has unveiled a major new artificial intelligence infrastructure initiative as the company accelerates efforts to secure long-term dominance in the rapidly intensifying AI race. Chief executive Mark Zuckerberg announced the launch of Meta Compute, a strategic program aimed at massively expanding the company’s data center capacity, energy usage, and underlying AI systems.

The move follows earlier disclosures that Meta planned to significantly increase capital expenditure to support its growing AI ambitions. Speaking publicly, Zuckerberg said the scale of infrastructure required to power next-generation AI products would redefine how technology companies compete over the coming decade.

According to the Meta CEO, the company intends to build AI infrastructure capable of consuming tens of gigawatts of power this decade, with plans to scale into the hundreds of gigawatts over time. Such levels of energy usage would place Meta among the world’s largest industrial power consumers, highlighting the immense computational demands of advanced AI models.

Zuckerberg said that how Meta designs, finances, and partners to deliver this infrastructure will become a core strategic advantage, shaping the company’s ability to deploy cutting-edge AI products at global scale.

Meta positions compute as the new competitive frontier

The announcement reflects a broader shift across the technology industry, where access to large-scale compute resources is increasingly viewed as essential for leadership in artificial intelligence. Training and deploying generative AI systems require enormous processing power, specialized chips, and globally distributed data centers capable of operating continuously.

Meta’s Chief Financial Officer Susan Li previously emphasized that AI infrastructure would be central to the company’s long-term strategy, arguing that control over compute capacity enables faster innovation and better user experiences. The launch of Meta Compute formalizes that vision into a dedicated initiative.

To lead the effort, Zuckerberg named senior infrastructure executive Santosh Janardhan as one of the project’s key architects. Janardhan will oversee technical architecture, software systems, custom silicon development, and the expansion of Meta’s global data center and network footprint.

Also playing a central role is Daniel Gross, who joined Meta last year after co-founding Safe Superintelligence alongside former OpenAI chief scientist Ilya Sutskever. Gross will lead long-term capacity planning, supplier partnerships, and strategic business modeling, focusing on how Meta scales compute resources sustainably and efficiently.

The emphasis on internal infrastructure suggests Meta is seeking greater independence from third-party cloud providers as it scales AI products across its platforms, including Facebook, Instagram, WhatsApp, and future mixed-reality offerings.

Governments, energy, and the future of AI scale

Recognizing the political and regulatory dimensions of such vast infrastructure expansion, Meta has also tasked senior executive Dina Powell McCormick with leading government engagement tied to the initiative. Her role will focus on working with governments and financial partners to support the construction, deployment, and financing of AI infrastructure projects.

As AI-driven data centers consume increasing amounts of electricity, energy availability has become a critical constraint. Industry estimates suggest AI-related power demand in the United States alone could increase tenfold over the next decade, raising concerns around grid capacity, sustainability, and environmental impact.

Meta’s announcement comes amid similar moves by major competitors. Microsoft has expanded partnerships with AI infrastructure providers, while Google parent Alphabet recently acquired a data center firm to secure additional capacity. The race to build generative-AI-ready cloud environments is now one of the most capital-intensive battles in the technology sector.

While Meta has not disclosed detailed timelines or locations for its planned infrastructure builds, the company’s ambitions signal a long-term commitment to AI at unprecedented scale. As compute power increasingly determines which firms can lead in artificial intelligence, Meta is betting that owning and operating massive infrastructure will be as decisive as the models themselves.