Bull & Bear

Bull and Bear

Verdict: Lean Long, Wait For Confirmation — the underlying ad auction is the most profitable in listed media and the FY25 multiple compression is real, but management has admitted it keeps underestimating compute needs and the depreciation cliff arithmetic is not yet visible in reported margins. The single tension that decides this name is whether AI capex converts into a wider auction spread before D&A from $69.7B FY25 plus $125–145B FY26 of PP&E lands in the P&L from 2027. The condition that would change the verdict in either direction is one quarter of operating cash flow growth visibly outrunning capex (bull) or a further upward revision to FY26/FY27 capex paired with the first negative-FCF print since 2022 (bear).

Bull Case

Bull-case fair-value anchor: ~$820. Method: normalized FY26 net income ~$85B (FY25 ex-OBBBA $76B grown ~12% on the +20% revenue trajectory and tax normalization to 13–16%) × 25× P/E (META's own 10-year median, below GOOGL's 29×) ÷ 2,574M diluted shares ≈ $825/share. This is the value implied if the bull-case inputs hold; treat it as scenario output, not a price forecast. Disconfirming signal: two consecutive quarters where impressions outrun price per ad (e.g., impressions +15% with price-per-ad flat or negative) while AI ad-tools run rate stops growing — that combination would mean the auction is monetizing via inventory dilution and the capex is not converting to pricing power.

Bear Case

Bear-case fair-value anchor: ~$480. Method: multiple compression to ~17× normalized earnings on ~$72B FY26 net income (ex-OBBBA tax cushion, ex-useful-life cushion of ~$3B). Cross-check: ~12× EV/EBITDA on FY26E EBITDA ~$120B (hyperscaler multiple, not software platform) ≈ $1.42T EV → ~$555/share; trough-cycle 14× P/E on $72B NI ≈ $400/share; $480 splits the two anchors. This is the value implied if the bear-case inputs hold; treat it as scenario output, not a price forecast. Cover signal: either (a) the Jan 2027 guide holds FY27 capex flat or trimmed versus FY26 actual with explicit ROI commentary; or (b) Meta discloses a direct AI revenue line above $15B annualized that converts the capex from sunk infrastructure into investment.

The Real Debate

Verdict

Lean Long, Wait For Confirmation. The bull carries more weight on the long arc because the underlying ad auction is accelerating (price per ad +10%, impressions +14% in Q3 25) on a 52%-margin Family of Apps base that has no listed peer within striking distance, and the balance sheet is still strong enough (net debt/EBITDA 0.02×, OCF +27% to $115.8B) to self-fund the AI build without breaking the equity story. The single most important tension is whether AI capex converts into a wider auction spread before incremental D&A from $69.7B FY25 plus $125–145B FY26 of PP&E lands in reported margins from 2027 — that is the durable thesis variable. The bear could still be right because management has openly admitted it keeps underestimating compute needs, the FY25 margin only held because of a one-time server-life extension, and the off-balance-sheet stack ($103.77B leases not yet commenced plus $131.05B other commitments) is roughly four times reported long-term debt and invisible to standard screens. The verdict flips to Lean Long if the FY26 operating margin holds at 40%+ with no further useful-life adjustments and operating cash flow growth visibly outruns capex by Q3/Q4 2026 — the durable thesis breaker. The verdict flips to Avoid if the Jan 2027 capex guide steps up again versus FY26 actual or the first negative-FCF quarter since 2022 prints without a disclosed AI revenue line above $15B annualized — the near-term evidence marker the market will trade on. Until one of those resolves, the franchise quality justifies attention but not commitment at the prevailing multiple.