Bill Ackman Says Stocks Are “Stupidly Cheap”

Key event: SpaceX is reportedly planning an IPO to raise about $75 billion at a potential $1.75 trillion valuation with up to 30% of shares earmarked for retail — a historic, valuation-driven offering that could be volatile if Starlink or Musk’s orbital AI/data-center ambitions underdeliver. Debate on AI: bullish view argues AI will deepen aggregator data moats and boost conversion for platforms like Expedia, Instacart and Uber; bearish view warns agentic AI could disintermediate third-party booking platforms. Ackman argues some high-quality names are 'stupidly cheap' and highlights Fannie Mae/Freddie Mac as long-term upside if conservatorship ends; hosts also flagged potential value ideas including Howard Hughes, Lululemon, Microsoft, Alphabet, Uber and SoFi.

Analysis

AI’s near-term impact on third‑party aggregators is a classic moat-versus-disintermediation trade: companies that own both the buyer interface and fulfillment (Uber, some Instacart-like models) can convert AI into higher conversion and lower marginal delivery cost, potentially boosting EBITDA margins by a few hundred basis points over 12–24 months. Pure demand‑aggregation plays without execution control (classic metasearch/travel sites) face a credible multi‑year risk from agentic assistants that can bypass UI layers, putting pressure on traffic, CPC yields and supplier negotiation leverage.

A second‑order hardware/cloud effect is underpriced: the orbital/AIdata narrative (and the general push to on‑prem/cloud TPU/GPU capacity) increases optionality for NVDA and cloud integrators (MSFT/GOOGL) but also raises cyclicality and funding competition. A mega IPO that soaks $50–75bn of primary capital and allocates ~30% to retail can temporarily rerate retail brokers and short‑term growth froth while draining IPO season liquidity from later‑stage private AI financings — expect episodic volatility in AI hardware names if retail flows chase headline IPO allocations.

On policy and value, the Ackman/Fannie‑Freddie/HHH thread is a long‑dated, government‑policy binary that leaves HHH and related real‑estate assets highly convex to legislative or Treasury moves; these are multi‑year, event‑driven opportunities, not tactical trades. Near‑term risks that flip these narratives are agentic AI implementation delays, regulatory pushback on data‑monetization, and a sharper than expected slowdown in consumer travel/retail spending that would compress multiples across LULU/EXPE and cyclically exposed names.

Key Decisions for Investors

• Long UBER (size 1.5–3% of liquid equity sleeve): buy shares on pullbacks to capture 6–12 month AI-driven margin expansion from routing/ETA optimization and increased delivery density; target 30–60% upside, hard stop at 18% below entry to limit execution risk.
• Pair trade — Long MSFT (1.5%) / Short EXPE (1.5%): express AI platform vs aggregator bifurcation over 6–18 months. Rationale: MSFT reaps infrastructure and enterprise AI spend while EXPE is more exposed to disintermediation; expect 2:1 relative outperformance if agentic adoption accelerates. Close on clear evidence of AI agent bookings replacing intermediaries (traffic share drop >10%).
• Defined‑risk long NVDA call spread (3–9 month expiry): buy a near‑the‑money call and sell an upper strike to fund premium; this captures continued GPU demand from cloud/AI projects (including any orbital/edge compute pilots) with capped downside. Aim for asymmetry ~3:1 reward/risk; reduce position if GPU order guidance misses by >15%.
• Small, activist‑convex position in HHH (0.5–1%): treat as a 1–3 year binary. Rationale: upside from conservatorship/asset re‑monetization is material (>2–3x) but policy and rate risk are real — keep position small, add only on confirmed regulatory/IPO filing milestones, and hedge with tail protection (buy 12–24 month puts equal to 25% notional) to limit downside.