Meta's AI Strategy Is Working -- But These 2 Red Flags Could Hold It Back

Meta is facing a rising AI capital burden, with 2025 capex up 84% to $72 billion and 2026 guidance raised to $125 billion-$145 billion from $115 billion-$135 billion. Revenue grew 22% in 2025 and 33% in Q1 2026, but costs rose 35% in the quarter, while Reality Labs lost $4 billion in Q1 2026, adding to profitability pressure. The article argues these investments could ultimately pay off, but the timing and return on capital remain uncertain.

Analysis

META’s issue is less that spending is high and more that the payback curve is lengthening while the core business is still being marked up as if AI monetization is immediate. That creates a near-term paradox: the better the ad product gets, the more management is incentivized to reinvest incremental returns into infrastructure rather than letting them flow through to FCF. If capex continues to outgrow revenue by a wide margin into the next 2-3 quarters, the market may start treating META less like a growth compounder and more like a utility with a permanently elevated reinvestment rate.

The second-order beneficiary is the AI supply chain, but not evenly. NVDA should keep capturing the largest share of incremental spend, while infrastructure-adjacent names with pricing power and bottleneck exposure are better positioned than application-layer winners. The real risk for META is that rising capex crowds out optionality: every dollar spent on training/inference hardware is a dollar not available for buybacks, which matters if growth normalizes and the multiple contracts before those investments mature.

Reality Labs is the bigger governance problem because it creates a hidden drag that can’t be easily optimized away in the near term. The market can tolerate one moonshot if the core engine is throwing off excess cash, but two simultaneous capital sinks raise the probability of a future capital-allocation reset. The contrarian angle is that the stock may already be pricing in a fair amount of skepticism; the catalyst for outperformance would be a clear deceleration in capex growth or evidence that AI spend is shifting from infrastructure buildout to monetization, which could re-rate margins quickly over 1-2 quarters.

For now, the setup favors being long the beneficiaries of META’s spend rather than the stock itself until visibility improves. If management signals another step-up in 2026 capex without a commensurate FCF inflection, the narrative shifts from ‘investing for growth’ to ‘structural margin compression,’ and that is when downside usually becomes more persistent.

Key Decisions for Investors

• Short META vs long NVDA into the next earnings cycle: use a 3-6 month horizon to express the view that incremental AI dollars accrue to the supplier with the most pricing power while META absorbs the capex drag; target a modest relative move rather than outright downside.
• Buy META put spreads 60-120 days out if the stock re-rates on ‘AI monetization’ headlines before capex guidance stabilizes; this is a defined-risk way to fade optimism while limiting theta bleed.
• Overweight NVDA and selective AI infrastructure winners over META for the next 1-2 quarters; the better risk/reward is with picks-and-shovels names where spend converts into revenue immediately rather than being capitalized internally.
• Avoid adding to META on strength unless management shows capex growth flattening sequentially; if capex stays above revenue growth for another quarter, expect multiple compression and weaker buyback support.
• For event-driven accounts, consider a pair trade long NFLX / short META only if advertising growth slows materially; that would isolate the market’s willingness to pay for monetizable engagement versus heavy internal reinvestment.