Meta Platforms (META)

Executive Summary & Action Plan

Verdict

HOLD — The best ad engine in the market just added ~$240B of market value in six July sessions for a compute-vendor option with zero disclosed revenue. We respect the pivot and won't pay for it in advance: fair value $592 sits 12% below spot, while the bull case ($786) stays live if Meta Compute lands customers.

Price $669.21 | Market cap $1.70T | Target $592 | Upside -11.5% IWANNAVY Fair Value · price as of 2026-07-11 close · Street consensus $827.91 (reference) — we are 28% below Street because Street keys off forward EPS frameworks blind to the 2026–28 FCF trough and hasn't re-marked since the 4/29 capex raise

Abstract

Meta's ad machine re-accelerated to +33.1% YoY on genuinely AI-driven economics — impressions +19%, pricing +12%, the value-optimization suite above a $20B run-rate — while operating margin held 40–43% through the spend surge. Both headline net-margin outliers are one OBBBA/CAMT tax whipsaw (−$15.93B charge in Q3'25, +$8.03B benefit in Q1'26); clean net income grew ~13%, not the optical +61%. The real event is capital allocation: the $125–145B capex guide is back-loaded (Q1 ran $19B; H2 needs ~$35–42B/quarter), flipping 2025's $46B FCF toward breakeven just as the market added ~15% of market cap in six sessions on Meta Compute (Bloomberg 7/1) and the Iris chip (Reuters 7/9) — a second P&L that has yet to bill a customer. Our $592 fair value is the log-median of a capex-charged DCF ($567), 15.5x forward EV/EBITDA ($786), and 26x normalized FCF ($592); our prior $827 mark (6/15) migrates to bull-adjacent now that the spend is committed and the revenue isn't. Rating: HOLD — reduce the July pop, re-enter at $592/$520.

Action Plan

Hold the core; take the July optionality-premium off into pre-print strength. Reduce $700/$760, re-enter $592 (FV) and $520; trader stop $495.

• Risk/Reward 0.53:1 at market | Prob-weighted 12M return −10.6% (Bull 25% × +17.5% + Base 45% × −11.5% + Bear 30% × −32.8%) | Confidence: Medium

---

Last Four Quarters

Revenue climbed $47.5B → $59.9B before Q1 seasonality ($56.3B, +33.1% YoY — the fastest of the Big-4). Gross margin is pinned at ~82%, operating margin in a 40–43% band. The two net-margin outliers are the same tax event in both directions: Q3'25's 5.3% carries the $15.93B OBBBA/CAMT valuation-allowance charge; Q1'26's 47.5% carries the $8.03B Treasury Notice 2026-7 partial reversal. Ex-items, quarterly net income runs $16.6–22.8B. FCF still healthy at ~$13B/quarter — because the capex ramp hasn't hit yet.

GAP verification (+5.97% on the 7/10 session): idiosyncratic and confirmed — Reuters reported (7/9, internal memo) the Broadcom-co-designed, TSMC-fabbed "Iris" accelerator enters production in September, anchoring the 7GW→14GW compute doubling and reduced Nvidia/AMD reliance; it compounded the 7/1 Bloomberg "Meta Compute" story (+9% that day). Two-step re-rating from capex spender to infrastructure owner-seller.

---

Revenue & Profit Mix

Advertising is ~97.7% of revenue; Family of Apps runs a 48.1% segment margin while Reality Labs burns ~$4B a quarter (cumulative losses >$90B). Q1'26 per the 4/29 8-K.

---

Business Lines

• Family of Apps — Advertising (~97.7% of revenue): 3.56B daily actives monetized by the GEM/Andromeda ranking stack and Advantage+. Must know — the re-acceleration is AI-attribution, not cycle: gen-AI creative advertisers doubled to ~8M and ad durability is now a function of model quality and the GPU fleet — exactly what the capex buys.
• Family of Apps — Other (~1.6%): WhatsApp Business, Meta Verified. Must know — WhatsApp is the portfolio's largest under-monetized asset, but it is a 2027+ story, immaterial to current numbers.
• Reality Labs (0.7% of revenue, −$4.03B/qtr): Must know — a permanent ~7pp drag on consolidated margin; any credible loss-cap commitment is instant EPS upside, and none was given on 4/29.
• Emerging: Meta Compute / Iris / Muse Spark hosting (0% of revenue): Must know — this is the entire July re-rating: a reported plan, pre-revenue, first Iris wafer due September; Wolfe estimates ~20% EPS uplift per monetized gigawatt.

---

IWANNAVY Fair Value

Log-median lands at $592. The EBITDA lens ($786) nearly reproduces Street because EBITDA is blind to the capex; the DCF ($567) charges ~$1.29T of 2026–35 capex at a 9.4% WACC. Our 6/15 fair value of $827 is retired: the compute-doubling spend is now committed while seller economics remain optionality — $827 is today's bull-adjacent case, not base.

• Thesis breaker (both directions): a named external Meta Compute customer or quantified Iris Nvidia-displacement on 7/29 forces our base toward $700+ (deep-bull $881); a third straight capex raise with no monetization detail and ad growth <25% re-opens the "two money pits" tape — deep-bear DCF $310.

---

Catalysts & Risks

---

References

• Meta Q1 2026 results — ads +33%, capex guide raised to $125–145B (8-K, 2026-04-29) · CNBC — Q1'26 report & $8.03B tax benefit (2026-04-29) · CNBC — Q3'25 $15.93B OBBBA tax charge (2025-10-29)
• Bloomberg — Meta planning cloud business to sell AI compute (2026-07-01) · Reuters — Iris AI chip into production in September (2026-07-09) · Wolfe Research — FY27 capex ~$220B estimate, must show non-ad revenue (2026-07-09)