Meta's AI Compute Monetization: Turning Capex into Revenue
News of Meta's potential move to sell AI computing power and model access to external clients has sparked diverse interpretations of its strategic direction. A recent analysis from Deutsche Bank sheds light on what this could really mean for the company's future.
Strategic Intent: Optimization, Not Retreat
The report argues that this initiative should not be seen as a pullback from cutting-edge AI or "super-intelligence" ambitions. Instead, it represents a pragmatic approach to asset utilization.
Meta is likely seeking to commercialize older, non-core, or intermittently idle computing infrastructure, while reserving its latest-generation, most powerful chips (like its next-gen in-house training accelerators) for internal, mission-critical model development. This is a classic tiered resource management play.
Shifting the Market Narrative
A persistent concern among investors has been the disconnect between Meta's massive capital expenditures and the perceived direct revenue returns, a gap widened by the AI arms race. Deutsche Bank suggests the cloud services move could pivot the market's focus.
The narrative may shift from scrutinizing "how much Meta is spending" to evaluating "how much Meta can earn" from these investments. This re-rating of the company's high-margin revenue optionality could positively impact its valuation framework.
The Multi-Billion Dollar Potential
Deutsche Bank's quantitative model projects that Meta's total AI-related compute capacity could reach 8 to 11.5 Gigawatts (GW) by the end of 2027. Within that, an estimated 1.2 to 2.7 GW of capacity could be made available for external sale.
The revenue projection is based on two key assumptions:
- Utilization Rate: 75% of the saleable capacity is successfully sold.
- Pricing Power: Each GW of compute generates $10 billion to $15 billion in annualized revenue.
Under this model, the business could contribute $9 billion to $30 billion in incremental annual revenue for Meta by 2027. This represents a 3% to 10% lift over the market's current consolidated revenue expectations for the company. The base-case scenario points to roughly $17.5 billion in new revenue.
This potential revenue stream is significant enough to be considered a substantive third growth pillar, diversifying beyond Meta's core advertising and in-app purchase engines.