Deutsche Bank: Meta's Cloud AI Offering Could Unlock $100 Billion+ Monetization
Deutsche Bank believes Meta's plan to offer AI computing power and model access to external clients will not compromise its cutting-edge AI development. Instead, the bank suggests Meta will likely monetize older, non-core, or temporarily idle computing resources while reserving its latest-generation chips for internal training. This strategic move could shift market perception from concerns about Meta's high capital expenditures and limited revenue returns to an appreciation of its new, high-margin revenue opportunities. Deutsche Bank estimates that by the end of 2027, Meta's AI-related computing capacity could reach 8 to 11.5 gigawatts (GW). Of this, approximately 1.2 to 2.7 GW could be made available for external sales. Assuming 75% of this capacity is sold, with an annualized revenue of $10 billion to $15 billion per GW, Meta could generate an additional $9 billion to $30 billion in revenue by 2027. This represents a 3% to 10% increase over current market revenue expectations, with a baseline scenario projecting around $17.5 billion in new revenue.
Deutsche Bank's assessment highlights a potential strategic pivot for Meta, aiming to leverage its substantial AI infrastructure investments for external revenue generation. This move could address market concerns regarding the significant capital expenditure associated with AI development by creating a new, potentially high-margin income stream. The analysis suggests a rationalization of resource allocation, where older or underutilized computing power is commercialized, thereby improving overall capital efficiency. This approach could position Meta to better navigate the evolving AI landscape, balancing internal innovation needs with external market opportunities. The long-term implications involve Meta's ability to scale this service effectively and manage the associated operational complexities, potentially setting a precedent for other large technology firms investing heavily in AI infrastructure.
AI-generated to prompt reflection — not editorial opinion, not advice, not a statement of fact. How this works.