Mark Zuckerberg Published His Superintelligence Manifesto Monday Morning. Meta Still Trades at 18 Times Forward Earnings.

Meta’s CEO Mark Zuckerberg pushed a “distribute superintelligence to everyone” agenda alongside a release of open-weight AI models (Muse Glimmer; Muse Spark 1.2 weights to follow), positioning distribution as both a safety plan and future compute monetization. The capital plan remains heavy: 2026 capex guidance lifted to $130B–$145B and Q2 capex was $31.1B, while free cash flow fell to $784M (from $12.4B in Q1) and net income decreased 14% YoY to $15.8B. Revenue grew 28% YoY to $60.8B, but costs rose 55% and third-quarter revenue guidance of $61B–$64B implies further moderation; shares were down ~10% in 2026 and ~25% from the 52-week high as the market questions the cost/revenue arithmetic.

Analysis

The market is likely mispricing where the economic moat sits. Open weights commoditize the model layer, but that is not where Meta ever intended to extract most value; the real lever is distribution, ad targeting, and incremental inference demand. That makes the bigger second-order winner the compute stack: every extra use case built on Meta’s models still consumes GPU-hours, networking, storage, and power, so the capex program effectively transfers value upstream even if the model itself is given away.

For META, the near-term risk is not ideology but capital intensity: free cash flow can stay suppressed for several quarters if management keeps revising the build-out higher faster than ad monetization improves. The stock can work only if investors believe AI improves ad ROI enough to re-accelerate revenue per impression within 1-2 quarters; otherwise multiple compression is the path of least resistance over the next 1-3 months. The falsifier is a pause in capex escalation or evidence that AI features are lifting conversion without proportional cost inflation.

NVDA remains the cleaner beneficiary on a 6-18 month view because Meta’s spend is a demand signal, not a licensing opportunity. The contrarian miss is that open-source rhetoric may actually expand total token consumption by lowering adoption friction, so the upside may accrue to picks-and-shovels rather than to model monetization. Netflix is mostly a bystander; any AI-driven ad-tech efficiency is too indirect to underwrite a trade here.

Key Decisions for Investors

• Long NVDA / short META into the next earnings/capex update: the pair expresses the view that AI spend converts into upstream hardware demand faster than it converts into near-term Meta FCF. Risk: META re-accelerates ad growth enough to re-rate the stock; exit if META restores FCF above ~$10B/quarter while holding capex flat.
• Use META weakness to sell 1-3 month call spreads rather than chase outright shorts: the stock still has ad-business support, but the capex overhang caps upside until the market sees durable monetization. Falsify if Q3 guidance or operating income inflects above expectations.
• Overweight semicap/AI infrastructure exposure via NVDA on any broad META-related tech selloff; the thesis is that open-weight releases increase inference usage and keep accelerator demand tight. Stop if hyperscaler capex commentary broadens into a 2027 slowdown.
• No trade in NFLX on this print; AI spillover into streaming ads is too indirect. Revisit only if Meta’s ad product gains visibly compress lower-funnel CPCs across digital ad peers.