Meta Platforms [META]
Task 6 · built 2026-07-27 · spot $593.87 · status: WATCHLIST (bearish bias, not actionable)
Ongoing maintenance hands off to equity-research:catalyst-calendar and equity-research:thesis-tracker.
Consensus: revenue ~$60.2bn (+26.6% YoY, 45 analysts); EPS $7.13–$7.23.
| Metric | Watch for | Threshold and action |
|---|---|---|
| FY2026 capex guidance | Currently $125–145bn (raised from $115–135bn on the Q1 call) | Raised again above $150bn → strengthens the bear case materially. CUT → strongly bullish, validates the option value; move toward Long |
| FY2027 capex indication | Not yet given | Above $180bn with FoA growth < 15% → trigger E1, the first short condition |
| FoA revenue growth | +33% Q1-2026, +26.6% consensus | < 20% → the deceleration is arriving faster than modelled |
| Total D&A | $6.0bn in Q1-2026 | > $8.0bn → the wall is arriving early; check against the model's $24.9bn FY2026E |
| Any useful-life commentary | 5.5 years since Jan-2025 | Any further EXTENSION → trigger E2. The single most informative possible disclosure |
| Reality Labs operating loss | −$19.2bn FY2025, still growing | > −$5.5bn in the quarter → the RL burn is still accelerating |
| Operating margin | 41% in Q1-2026, down from a 48% Q4-2024 peak | < 38% → margin compression is running ahead of the model |
| Effective tax rate | Distorted by CAMT in both directions | Normalise before comparing to consensus. Q1-2026 EPS of $10.44 included an $8.03bn benefit worth $3.13/share |
| DAP | 3.56bn, +4%, sequentially DOWN on Iran/Russia disruptions | A second consecutive sequential decline would be the first genuine user-base signal in years |
IV context: 21-Aug ATM IV is 52.8% against 38.8% realised — 1.36x. The market prices an 11.0% move. No position is taken into this print because the house has no view that the move exceeds what is priced.
The Hyperion VIE disclosure. Track the maximum-exposure-to-loss figure each quarter: $45.95bn (Dec-25) → $45.99bn (Mar-26). Trigger E3: any move of the RVG from "not probable" to a recorded liability, any disclosed impairment at the Venture, or a step-change in maximum exposure. This is a footnote most models never read and it is where the structural risk actually lives.
Prometheus (New Albany, Ohio, 1 GW, gas-powered) is scheduled to come online in 2026. Trigger: the first quarter in which Prometheus depreciates is the first real-world test of the 5.5-year life against actual utilisation. Hyperion does not operate until 2029 — a reminder that the largest single project contributes nothing to the P&L within the forecast horizon.
Six show-cause orders to all RTOs. Meta is more exposed than Amazon here because it has publicly committed to specific gigawatt figures at named sites with named utilities (Entergy). Trigger: any ruling restricting co-located or behind-the-meter load at Richland Parish or New Albany directly threatens a dated, disclosed schedule.
Same grid, plus: whether FY2027 capex has been formally guided, and whether consensus FY2027 EPS has moved above this model's $35.46 (trigger E4 — the only route by which a genuine negative variant could come into existence).
This is the stated reassessment date for this Watchlist entry, and the single most important scheduled event on the name. It carries: - The annual useful-life study. Meta extended to 5.5 years in January 2025; Amazon shortened in the same month. A third data point on Meta's assumption is what converts this from a published thesis into a falsifiable one. Either direction is highly informative. - The updated contractual-commitments ladder (was $131.0bn) — watch whether it converges toward, or stays far below, the capex run-rate. The gap is the option value. - Full-year Hyperion VIE disclosure, RVG threshold decay, and lease-commencement detail. - Any first disclosure of AI-attributable revenue.
RL has lost $52.9bn over three years and losses are still growing, while the word metaverse has been absent
from every earnings release for five consecutive quarters. Trigger: any announced restructuring, wind-down or
material spending cut at RL would add ~$19bn of annualised operating income and is a live bullish catalyst that
no bear case should ignore.
Modelled D&A: $24.9bn (FY2026) → $39.2bn (FY2027) → $55.9bn (FY2028) → $74.1bn (FY2029). Modelled operating margin: 41.1% → 36.8% → 32.6% → 29.1%. The crossover — when D&A growth exceeds EBITDA growth — is modelled at FY2027. Watch each quarter for it arriving earlier. This is arithmetic, not forecast.
The implied-penetration requirement: annual revenue per daily active person must rise from ~$56 to ~$88 by FY2029 (+57%), equivalently Meta's share of global digital advertising from ~22–24% to ~30–32%. Checkable annually from the disclosed DAP and revenue figures — a rare case where a TAM claim has a directly reported denominator.
Meta raised $29.9bn net in November 2025. Trigger: any subsequent issuance at a materially wider spread signals the credit market repricing hyperscaler capex risk — a sector-level early-warning indicator that would arrive before the equity market reacts.
| Trigger | Date | Effect |
|---|---|---|
| E1 — FY2027 capex > $180bn with FoA growth < 15% | Q2/Q3-2026, then 10-K | First short condition |
| E2 — Any further extension of server useful life | FY2026 10-K, ~Jan-2027 | Second short condition — the decisive one |
| E3 — RVG recorded as a liability / Venture impairment / max-exposure step-change | Any 10-Q Note 5 | Third short condition |
| E4 — Consensus FY2027 EPS rising above $35.46 | rolling | Fourth short condition — creates the missing variant |
| T1 — tape condition | ALREADY SATISFIED (12-1 momentum −22.5%, −6.7% vs 200dma) | — |
| L1 — price ≤ $430 with FoA growth ≥ 15% | rolling | Long conversion |
| L2 — a capex CUT | Q2/Q3-2026 or the 10-K | Long conversion — validates the option value |
| L3 — Reality Labs restructuring or wind-down | undated | Long conversion — ~$19bn of annualised operating income |
| L4 — any disclosed, separable AI revenue line | undated | Long conversion — makes the programme measurable |
Short conversion requires E1, E2, E3 or E4. Unusually, the tape condition T1 is already met — which is precisely why the discipline matters here: a favourable tape is the easiest thing to mistake for a thesis.