$META is preparing to lease its excess AI infrastructure capacity. The core operational and financial drivers behind this decision include:

Offsetting Capex: With annual AI capital expenditures reaching an estimated $125B–$145B, renting GPU capacity creates a direct B2B revenue stream to help offset these infrastructure costs.

Monetizing Idle Compute: Rather than leaving surplus computing power unused between internal model training cycles, Meta can commercialize the excess hardware capacity.

Cloud Diversification: This move marks Meta's entry into the infrastructure-as-a-service market, positioning it alongside traditional cloud providers like AWS $AMZN , Google Cloud $GOOG , and Azure $MSFT
Ultimately, the strategy leverages Meta's physical asset base to diversify its revenue model beyond digital advertising and capture market demand for raw compute power.