Meta's Earnings Got a Major Tax Boost. Here Are the Adjusted Figures You Need to See.

Meta reported Q1 2026 revenue of $56.3B, up 33% year over year and ahead of the prior quarter’s 24% growth, but adjusted EPS rose only about 14% after excluding an $8.03B tax benefit. Operating income increased 30% to nearly $23B while total expenses climbed 35%, highlighting margin pressure from rising costs. Management also raised 2026 capex guidance to $125B-$145B from $115B-$135B, signaling materially heavier AI and infrastructure spending.

Analysis

META is still compounding, but the quality of compounding is deteriorating: revenue acceleration is increasingly being translated into infrastructure intensity rather than shareholder earnings. That usually marks a mid-cycle transition where the stock stops trading on top-line scarcity and starts trading on whether incremental ad monetization can outrun depreciation, cloud commitments, and AI capex creep. The market is likely underestimating how much of the next 12-24 months of headline growth is already prepaid through future fixed-cost obligations.

The second-order winner is NVDA, but not in a straight-line way: Meta’s larger capex envelope supports demand visibility for accelerators, networking, and memory, while also tightening supply for other buyers and keeping pricing firmer across the AI stack. The loser set is more subtle — every large-cap platform chasing inference/training capacity raises the hurdle rate for its own equity holders because the industry is collectively moving from software-like margins toward utility-like depreciation schedules. That is why seemingly strong operating margins can coexist with flattening EPS leverage.

The key risk is that consensus is anchoring on the ad business reacceleration and ignoring the lagged margin compression from contractual commitments. Over the next few quarters, the debate should shift from revenue growth to free-cash-flow conversion and incremental ROIC on AI spend; if those metrics disappoint, the stock can re-rate lower even with healthy ad trends. Conversely, a visible moderation in capex growth or evidence that custom silicon meaningfully reduces unit costs would be the main catalyst to re-rate the name higher.

The contrarian take is that the stock may not be expensive on near-term earnings, but it can still be expensive on normalized cash flow if capex settles at a structurally higher plateau. That creates an asymmetric setup where the business can keep beating revenue while the equity underperforms because expectations for future cash generation are being diluted by multi-year investment intensity.

Key Decisions for Investors

• Maintain a tactical underweight in META vs. other mega-cap platforms for the next 3-6 months; the risk/reward is capped until the market gets proof that 2026 capex is converting into higher FCF rather than just higher depreciation.
• Express the AI-infrastructure beneficiary trade via long NVDA on pullbacks, but hedge with a basket short of mega-cap capex-intensive software/platform names; the theme is demand visibility into the supply chain, not enthusiasm for all AI spend.
• Consider a pair trade: long NVDA / short META into the next earnings cycle, targeting upside from infrastructure demand while fading the probability that Meta’s incremental spend immediately improves equity returns. Use a tighter stop if META signals capex moderation.
• For investors already long META, overwrite with near-dated calls or trim into strength; the near-term upside is likely more limited than the downside if guidance implies another step-up in capex or contractual commitments.
• Watch for a catalyst on custom silicon efficiency over the next 1-2 quarters; if unit economics improve enough to stabilize FCF margins, cover shorts quickly because the re-rating could be sharp.