Meta has announced Meta Compute, a new top-level initiative aimed at building AI infrastructure at unprecedented scale. The effort is designed to ensure that compute capacity does not become a bottleneck as AI systems continue to advance.

The initiative will be co-led by Meta’s infrastructure chief Santosh Janardhan and Daniel Gross, who joined Meta last year from AI safety startup SSI. Newly appointed Meta president and former U.S. national security official Dina Powell McCormick will oversee government partnerships to help finance and build large-scale capacity.

Meta has already committed approximately $600 billion in U.S. infrastructure spending by 2028 and has recently secured long-term nuclear power agreements to supply energy for its data centers, signaling how central compute and energy access have become to AI strategy.

The announcement arrives alongside reported layoffs in Meta’s Reality Labs and metaverse/VR divisions, with roughly a 10% reduction expected, underscoring a strategic reallocation of resources toward AI infrastructure.

Why it matters: As the AI race increasingly shifts from model development to infrastructure dominance, Meta is doubling down on compute and energy at nation-scale levels. The move reflects a broader industry reality that access to massive, reliable compute may define who can reach and sustain the AI frontier.