Meta shares drop as Zuckerberg plans $145bn AI spending spree

Meta shares fell more than 6.5% after it raised 2025 capital expenditure guidance from up to $135bn to up to $145bn, signaling another $10bn of AI/data-center spending. The company posted solid results with revenue up 33% to $56.3bn and profit up 61% to $26.8bn, but investors focused on the heavier AI spending burden and looming cuts of up to 8,000 staff. The article also highlights broader pressure on AI-related megacap tech stocks, with Microsoft and Amazon down about 1% in after-hours trading.

Analysis

The market is starting to distinguish between AI monetization and AI capex intensity. Google is being rewarded because incremental spending is still legible as operating leverage, while Meta is being punished because management is asking investors to underwrite a much larger, less transparent payback horizon; that creates a higher duration asset profile at exactly the moment rates, valuation discipline, and AI skepticism are tightening. The second-order effect is that capital allocation credibility matters more than raw AI ambition — a company can beat on earnings and still de-rate if investors think the marginal dollar of spend is less productive than peers.

This also sets up a relative scarcity trade across the AI ecosystem. If Meta is forced to spend above plan to catch up, its demand for compute, networking, and power infrastructure likely remains structurally high, but the incremental beneficiary is not necessarily META equity holders; it could be suppliers and infrastructure owners that monetize the arms race regardless of end-user ROI. Conversely, if OpenAI growth is even modestly below the market’s implied trajectory, the whole private-to-public AI spending complex becomes vulnerable to a multiple reset over the next 1-3 quarters, with the sharpest pressure on names whose valuation embeds perpetual hypergrowth.

The near-term risk is less about the quarter and more about guidance credibility into year-end and 2026. If capex keeps rising without a clear step-up in AI revenue per user or per workload, META can stay under pressure for months; if management can show measurable payback in ad ranking, engagement, or AI assistant monetization, the stock can recover quickly because sentiment is now extremely washed out. GOOGL looks like the cleaner contrarian long: the market is rewarding evidence that AI spend is translating into broad-based business lift, and that narrative can persist unless spending begins to outrun cash generation materially.

The selloff in META may be overdone tactically, but not strategically — the stock likely trades as a narrative battleground until the next capex update. The bigger overlooked risk is that this pullback could spread from single-name disappointment into a broader repricing of AI winners versus AI capex financiers, especially if Microsoft and Oracle continue to signal aggressive investment without proportional monetization. That would favor a dispersion trade rather than a broad index call.

Key Decisions for Investors

• Long GOOGL / short META into the next 4-8 weeks: own the name where AI spend is being re-rated as productive and fade the company asking for a much larger funding envelope with weaker investor trust; target a 10-15% relative move if capex concerns persist.
• Initiate a tactical long in META only after the first post-earnings stabilization day, using defined risk via 1-2 month call spreads or a partial position; the setup is for a reflexive rebound, but only if management can anchor payback metrics.
• Short ORCL on strength over the next 1-3 months: it remains exposed to the market’s growing intolerance for AI capex without immediate monetization, and the name is more vulnerable if AI enthusiasm cools further.
• Pair long MSFT / short META for a 1-2 quarter horizon: MSFT has the cleaner narrative on AI monetization and a more credible capital return framework, while META faces the highest scrutiny on spend discipline.
• Buy NVDA only on a broader AI-capex pullback, not on this headline alone; use weakness as an entry if the market overreacts to demand concerns, since the supply chain still benefits from every additional dollar of infrastructure spend.