Zuckerberg Is Quietly Turning Meta Into A Compute Provider Ahead Of Q2 Earnings

Meta’s stock is no longer trading only on ad efficiency; the real multiple driver is whether incremental AI capex can be recast as a revenue-bearing asset. If management can show third-party compute demand or even credible near-term utilization, the market should start capitalizing part of the depreciation stream more like infrastructure than pure growth spend, which can narrow the valuation discount quickly.

The second-order winners are the picks-and-shovels beneficiaries that keep the capacity full: NVDA, ANET, and power/cooling suppliers, because any proof of leased compute implies sustained training/inference intensity rather than a one-off buildout. The loser set is less obvious: hyperscalers such as AMZN, MSFT, and GOOGL could face small but real narrative pressure if a consumer-platform company demonstrates that idle capacity can be monetized outside the traditional cloud oligopoly.

Near term, the risk is that investors focus on capex optics and ignore utilization. Over the next 1-2 quarters, absent disclosed bookings or AI-driven ad ROI, this still reads as a cash-flow timing issue and can keep the stock range-bound. Over 6-18 months, wearables and model integration matter more: Meta Glasses is not a P&L driver today, but if it lifts engagement and ad inventory, it becomes an option on a new distribution layer. The consensus may be underpricing how much a credible non-ad revenue stream changes the sum-of-the-parts debate, even if margins are initially lower than core ads.