Why Meta’s stock could see a 50% rally, thanks to an overlooked AI wild card

Meta’s stock is down 6% over the past month, but Evercore ISI’s Mark Mahaney flags an AI-related opportunity: a potential $22B annual gross revenue run-rate starting in 2027. The bullish setup hinges on an ongoing compute shortage that could let Meta monetize capacity at a premium, turning elevated AI spending into a revenue tailwind. The article frames this as an “overlooked AI wild card” that could support a potential ~50% rally.

Analysis

The market is still pricing Meta’s AI spend as a one-way margin tax, but the more interesting setup is optionality: if compute remains scarce, fixed depreciation can be turned into a high-margin revenue stream rather than a pure expense. That’s the kind of asset-light monetization story that can rerate a stock even before the dollars show up, because investors typically undervalue 2-3 year revenue paths until there is a visible pilot or disclosure.

The first-order winners of persistent compute tightness are still the chip and infrastructure layer, not the platform consuming the compute. Meta’s edge is different: it has enough scale to amortize AI costs across a massive ad engine, so any improvement in ranking, targeting, or AI-assisted ad creation has a much bigger P&L lever than for smaller ad platforms. That makes SNAP and PINS the cleaner relative shorts if the thesis is “AI spend hurts everyone,” because they get the cost inflation without the same monetization upside.

The key risk is timing. In the next 1-3 earnings cycles, the stock will trade less on the 2027 monetization story and more on whether capex guidance keeps stepping up without evidence of revenue leverage; that would re-open multiple compression. The bull case is falsified if Meta shows no external compute revenue, no ad-yield improvement, and no utilization disclosure by 2026, because then the market will treat this as another expensive platform experiment rather than a new earnings stream.

Key Decisions for Investors

• Buy META on weakness via a 6-12 month call spread rather than stock if you want to express the 2027 optionality; the risk/reward works only if the market keeps underpricing future monetization. Invalidate if next earnings raise capex again with no mention of revenue conversion or utilization.
• Pair trade: long META / short SNAP + PINS basket for the next 3-6 months. Thesis is that Meta can spread AI costs across a far larger ad base and potentially improve ad yield, while smaller ad platforms absorb the same AI cost inflation with less pricing power.
• Do not chase the headline into NVDA/AMD here; they are the more direct beneficiaries of sustained compute scarcity. Use META as a watch item for any disclosure of external AI services or capacity monetization, then rotate into semis on pullbacks if that catalyst appears.
• Set an alert on META for the next earnings capex guide and any commentary on AI monetization. If the company does not show a path to incremental revenue within 1-2 quarters, reduce longs because the multiple can re-rate lower quickly.