How to Buy Pre-IPO Shares in Anthropic and OpenAI (And Why You Might Not Want To)

OpenAI and Anthropic filed confidential SEC draft S-1s, with IPO timing likely in late 2026/early 2027, valuing OpenAI at $852B (target ~$1T; ~50x $20B annualized run-rate) and Anthropic at $965B (target ~$1T; ~111x $9B run-rate). The article warns that only 5%–10% of shares will be floated, potentially driving sharp post-IPO price spikes, but pre-IPO exposure via secondary marketplaces/SPVs may be voided by the companies and become worthless. It also cautions that mutual funds/ETFs using proxy SPV exposure could trade far above NAV ahead of listing and then drop when investors switch into the actual public shares.

Analysis

The public-market edge here is not in owning the private AI winners; it is in understanding who monetizes the demand for access. Secondary brokers and SPV wrappers can see short-term flow spikes, but the economics are fragile because issuers can shut the tap at any time and buyers are mostly trading story, not verifiable cash flow. That creates a classic setup where the intermediary makes fees while the end investor takes the mark-to-model risk.

The more interesting second-order effect is rotation. If late-2026/2027 IPOs land with tiny float and big opening pops, the likely post-listing trade is out of proxy vehicles and into the actual stock, which should pressure any fund whose performance is dominated by private-AI markups. That argues for caution on proxy wrappers trading at a premium to NAV; the catalyst is not the IPO itself, but the moment public shares become available and the premium collapses.

Contrarianly, the market may be underestimating how little of this is investable alpha versus a timing game. The real winners over the next 12 months are probably the infrastructure names funding the capex cycle, not the access vehicles: the proxies can re-rate on sentiment, but they have no operating leverage to AI adoption. The thesis breaks if listings are delayed materially, secondary transfer restrictions are loosened, or proxy funds can prove durable NAV support independent of the IPO narrative.

Key Decisions for Investors

• Avoid buying ARK Venture Fund (ARKVX/ARKVX-like wrappers) or VCX at a premium to NAV; use any retail-driven premium expansion as an opportunity to fade the trade over a 3-12 month horizon. Falsifier: premium stays sticky through the first filing wave and remains supported after the first lockup/secondary events.
• Short SPCX or other public proxy vehicles on any spike tied to OpenAI/Anthropic hype, targeting NAV compression into the actual IPO window. Risk/reward is best if you can enter when the premium to reported NAV widens materially; stop if the premium fails to expand after the next fundraising/news cycle.
• Watch HOOD and FRGE as secondary-market flow beneficiaries, but only as tactical longs if reported volumes and fee revenue inflect. Otherwise treat them as limited-duration sentiment trades, not structural AI winners. Falsifier: no pickup in secondary transaction volume over the next 1-2 quarters.
• Prefer the AI infrastructure basket over private-AI proxies: maintain or add to NVDA on weakness relative to proxy-fund sentiment, since capex spend is still the cleaner monetizable stream over the next 6-18 months.