Will OpenAI or Anthropic IPO first?

**Stage 1 — Blind analysis (ignore market price):**

The most important factual inputs are: Anthropic has reportedly confidentially submitted a draft S‑1 to the SEC (the standard early step toward an IPO), while OpenAI’s executives publicly emphasize *flexibility* and have not committed to an IPO timeline. On a long horizon to 2040 (≈14 years), the default baseline for high-growth, high-profile private companies with deep VC and public-market investor interest is that they will pursue public liquidity options at some point unless there are strong structural or strategic disincentives.

Key structural differences matter: OpenAI’s unusual governance (capped‑profit model, partnership with Microsoft which holds a large economic stake and strategic control over cloud infrastructure) creates additional friction for a fast, straightforward IPO — it has both reasons to delay and concrete alternatives (direct secondary sales, continued private financing via Microsoft, or an acquisition/privatization arrangement). By contrast, Anthropic appears to be on a more conventional corporate footing and has signalled the canonical step toward listing: a confidential S‑1 filing. Confidential S‑1s are a strong signal — they are the usual step before an IPO marketing process, even though they do not guarantee a public filing or an immediate offering date.

Weighing incentives: investors and employees in both firms have strong incentives for liquidity; the macro environment (capital markets, appetite for tech AI names) will modulate timing but is unlikely to eliminate the probability of at least one IPO before 2040. Regulatory risk (antitrust, AI-specific regulation), potential acquisitions (Microsoft or other tech giants), or strategic decisions to remain private are real but, across the two companies and a long timeframe, unlikely to prevent both from ever going public. Given Anthropic’s confidential filing and OpenAI’s hesitancy, the posterior likelihood that *one of them — and almost certainly Anthropic — lists before OpenAI* is substantially greater than chance.

Putting numbers on it (blind): - Anthropic completes an IPO before 2040: ~85–92% (confidential S‑1 materially raises probability but does not guarantee success). - OpenAI completes an IPO before 2040: ~70–85% (structural complexities and Microsoft stake reduce but do not remove probability). - At least one of them IPOs before 2040 (the event as framed): combining the two and accounting for correlation I estimate **~92%**.

**Stage 2 — Market calibration (incorporate current market price):**

Current market price: Yes = 0.77. My independent 0.92 is meaningfully higher than the market consensus. Possible reasons the market prices this lower:

- **Ambiguity in question interpretation.** Some traders may read the market question as *“Which of the two will IPO first?”* (a comparative question) or as requiring a specific company to IPO first, rather than the simpler event *“At least one will IPO before 2040.”* Ambiguity reduces price. - **Conservative discounting of confidential filings.** Many market participants downweight confidential S‑1s because filings can be delayed, withdrawn, or converted into alternative outcomes (ACQ, direct listing, extended review), so they assign lower probability than I do to a completed IPO before 2040. - **Tail risks priced explicitly.** Traders may be focusing on the non‑negligible outcomes that block IPOs: prolonged regulatory action (national security/AI rules), a strategic acquisition (e.g., Microsoft buys Anthropic outright), or a corporate decision to remain private for control reasons. Those outcomes are possible but, spread over both companies across 14 years, I see them as lower probability than market implies. - **Liquidity/market composition effects.** This market has substantial volume (~228k contracts). Large participants may be hedging correlated bets elsewhere (e.g., other AI IPO contracts) or using this contract to express complex alternatives, pulling the price toward a conservative midpoint.

Conclusion: I view the market as somewhat underpricing the straightforward reading of the event (that at least one of OpenAI/Anthropic will IPO before 2040). If the event wording indeed means "will either OpenAI or Anthropic IPO before 2040?", then the market's 77% is too low relative to the objective signals (confidential S‑1 + long time horizon). If the question is interpreted differently by traders, that could explain the gap. Absent reinterpretation, the market appears to be underestimating the combination of Anthropic's filing momentum and the long horizon.

Catalysts that would rapidly move prices toward my view: - Public announcement of an effective S‑1 or an IPO filing date for Anthropic. - Reports of OpenAI board decisions explicitly ruling out early IPO (would lower chance OpenAI is first but not change "at least one"). - Acquisition rumors or binding deals (Microsoft/others) that would remove one company from the IPO path (would push price down).