Zuckerberg's Insane AI Spending Could Become a Windfall, If Meta Copies SpaceX's Playbook

Meta’s full-year 2025 capex surged to $69.7B from $37.3B and its 2026 capex range was raised to $125B–$145B, but Q1 revenue still jumped 33% to $56.3B with a 41% operating margin as ad impressions (+19%) and price per ad (+12%) rose together. The article highlights downside risk from capex/ROI concerns—Meta reportedly burned 60.2% of operating cash flow on capex in 2025 and Reality Labs still loses about $4B/quarter—tempered by a leaked compute-externalization plan and existing large infrastructure commitments (e.g., $107B in multiyear deals this quarter). Overall, it frames $593 as an entry for upside toward the $828.17 Street target (~39.8%), implying a re-rating if compute-as-a-service gains anchor tenants.

Analysis

The market is still treating Meta’s AI buildout like an undifferentiated cost center, but the more important mechanism is balance-sheet optionality: if even a small slice of internal capacity is monetized externally, the depreciation load gets amortized across a broader revenue base and the multiple can re-rate closer to infrastructure names than ad-platform peers. That matters most over the next 1-3 quarters, because the stock’s biggest driver will be whether management can point to signed third-party demand rather than just internal usage and promises.

Second-order winners are the companies that monetize the picks-and-shovels layer, not the obvious ad beneficiaries. AVGO is the cleanest fundamental lever if Meta’s custom silicon cadence stays intact; AAOI is more of a sentiment beta and likely overreacts to every capex headline, so it is a trading vehicle, not a thesis. The loser set is less about named competitors and more about other AI spenders: if Meta proves it can underwrite hyperscale economics from an ad cash engine, it raises the bar for every software or cloud platform claiming AI infrastructure ROI.

The contrarian miss is that the downside may already be mostly in the equity, while the upside requires only one credible external-compute anchor tenant to shift the narrative. The real falsifier is not the existence of capex; it is a sequence of quarters where capex stays elevated, no outside revenue appears, and operating losses elsewhere widen enough to turn the AI project into pure EPS dilution. That failure mode would show up first in forward margin guidance and free-cash-flow conversion, not in the headline revenue line.

Key Decisions for Investors

• Buy META on further weakness or hold/add on any pullback toward the prior support zone; 1-3 month thesis is that the stock is pricing capex as permanent expense while underpricing monetization optionality. Invalidate if Q2 revenue misses guide or management raises capex without any incremental third-party demand commentary.
• Structure a 3-6 month META call spread rather than outright shares if you want convexity into a year-end external-compute announcement; the trade works if one anchor tenant is disclosed, but premium should be the maximum loss if contracts remain absent.
• Pair long META / short AAOI for a 4-8 week trade around AI infrastructure sentiment. AAOI has high beta to capex headlines but little visibility into durable revenue capture, while META owns the underlying demand engine and optionality.
• Add AVGO selectively on any confirmation that Meta’s custom silicon roadmap remains on track; this is a slower 6-18 month beneficiary trade tied to design wins and content per accelerator. Exit if Meta shifts budget meaningfully away from in-house ASICs toward external GPU sourcing.
• Do not short META purely on capex size; if you need a bearish expression, wait for two consecutive quarters of no external compute revenue and weaker margin guidance, then use a short/META-vs-software basket rather than a standalone short.