Want to Own Stock in SpaceX, OpenAI, and Anthropic Pre-IPO? Here's How.

ARK Venture Fund holds concentrated exposure to private AI and aerospace leaders (SpaceX 17%, OpenAI 11%, Anthropic 4%) and charges a 3.49% gross expense ratio as an interval fund with quarterly repurchases. SpaceX posted roughly $16B revenue and $8B profit last year, had a Dec valuation of ~$800B and — after merging with xAI — is reportedly targeting an IPO valuation up to $2T. OpenAI’s ARR topped $25B in Feb 2026 (up 17% from end-2025), with a $852B post-money valuation and guidance to reach ~$175B revenue by 2029 and profitability by 2030. Anthropic’s ARR reached ~$30B in Apr 2026 (up >200% from end-2025), carries a $380B post-money valuation, projects ~$150B revenue by 2029, and expects profitability by 2028.

Analysis

The coming wave of high-profile private AI and aerospace issuances will reallocate economics across the stack rather than simply boosting a single beneficiary. Hyperscalers stand to capture recurring model‑hosting and data‑ingestion rents (high margin, sticky) while silicon vendors capture front‑end capex; that bifurcation implies materially different revenue cadence and multiple re‑ratings over 12–36 months.

A concentrated, illiquid private exposure product creates path‑dependent pricing events: quarterly liquidity windows and wide mark‑to‑model spreads can produce abrupt NAV gaps if sentiment shifts or an IPO disappoints, amplifying forced selling into public cleantech and AI names. That dynamic favors liquid, low‑beta ways to harvest the private-to-public repricing and penalizes retail channels that distribute illiquid exposure without explicit liquidity premia.

Key tail risks include faster-than-expected model efficiency (reducing incremental compute demand), tougher antitrust/regulatory scrutiny on dominant model owners, and slower enterprise monetization — any of which would compress multiples across both cloud and chip suppliers over quarters not years. Conversely, durable enterprise adoption or a surprise monetization channel (e.g., advertising or vertical SaaS embeds) would rapidly re‑rate hyperscalers and incumbents that control data plumbing.

My tactical read: prefer bifurcated exposure—own cloud infra optionality at scale while using pair/option structures to hedge semiconductor multiple risk. Monitor IPO cadence and first‑quarter aftermarket performance of any listed AI platform as a catalyst; treat fund‑level liquidity windows as local volatility multipliers rather than sources of alpha.

Key Decisions for Investors

• Buy MSFT Jan‑2027 call spread (buy 1yr 10% ITM calls, sell 25% OTM) to capture enterprise AI hosting tailwinds. Target: +30–50% if cloud AI revenue acceleration hits consensus; max loss = premium (~4–6% of notional). Timeframe: 9–15 months.
• Buy AMZN Jan‑2027 calls (or 1x/2x call spread) to ride AWS model‑hosting capture and advertising/commerce synergies. Target: +40% on 12‑18 month horizon if AWS maintains >10% incremental revenue contribution from model hosting; downside = option premium (limit to 3–5% portfolio risk).
• Pair trade (3–6 months): long AMZN or MSFT equal‑dollar / short NVDA equal‑dollar to exploit potential NVDA multiple compression if model efficiency or alternative accelerators emerge. Target pair return 15–25%; cut if NVDA outperforms relative leg by 20% (stop‑loss).
• Overweight SCHW for 6–12 months as a liquidity/flows play — RIAs and platforms reallocate into liquid ways to access private AI exposure. Target +15–25% upside from fee capture and incremental advisory flows; set stop at −10% to limit exposure to fee‑compression shock.