Meta Platforms: AI Return on Invested Capital Is Uncertain (NASDAQ:META)

The market should focus less on AI narrative upside and more on capital-intensity risk: when a company starts looking for third-party demand to absorb spare compute, it is effectively admitting depreciation is outrunning monetization. That is a free-cash-flow issue, not an accounting one, and it can compress META’s multiple even if ad revenue stays healthy. Near term, the spend remains a clear demand tailwind for NVDA and the broader AI supply chain, but that benefit is mechanical and likely better reflected in suppliers than in the buyer.

The 1-3 month catalyst is disclosure quality. Bulls need evidence that AI-driven ad efficiency is lifting revenue per user or auction pricing enough to cover the incremental capex, and that external leasing is additive rather than a symptom of overcapacity. If management keeps deflecting ROIC questions, the stock can de-rate on a clean print because investors will start capitalizing META more like an infra-heavy platform than a capital-light ad compounder.

Contrarian view: consensus is still treating this as a benign reinvestment cycle, but the risk is that AI spend crowds out the buyback support that has historically underwritten META’s per-share compounding. The thesis breaks if Meta shows a measurable step-up in operating leverage or gives a credible path for lease revenue to offset depreciation; absent that, the burden of proof stays on bulls. This is less a near-term earnings miss story than a 6-18 month quality-of-earnings debate.