Tools chosen: polymarket_direct polymarket_related kalshi_related claude_news gdelt_news wikipedia code_execution

Sub-questions (Fermi decomposition)

Planner reasoning

This is a conjunctive question: OpenAI must (a) actually complete an IPO with a first trading day on or before Dec 31, 2027, and (b) close that day above $1T market cap. The key research is the current Polymarket price anchor, the state of OpenAI's IPO preparations/restructuring, its latest private/secondary valuation trajectory, and the ladder of other threshold markets (e.g. 'OpenAI IPO in 2026/2027?', '>$500B', '>$2T') which lets me back out the conditional probability of the $1T level given an IPO.

# Current state OpenAI has not IPO'd yet. It confidentially filed a draft S-1 with the SEC in June 2026, but as of mid-August 2026 no exchange, ticker, date, or pricing has been set. CFO Sarah Friar is pushing internally for a 2027 listing (citing readiness/reporting-standards concerns) while Sam Altman reportedly prefers Q4 2026 and has called any valuation below $1T a "nonstarter." Most recent private-market reference price is $852B (Aug 2026 employee buyback), below the $1T IPO target. # Timeline of key events - 2025 (Oct 27, confirmed): OpenAI completes nonprofit→PBC restructuring; Microsoft deal gives it ~27% stake; OpenAI valued ~$500B in that transaction (Wikipedia/Yahoo/TipRanks). - 2026 (early, confirmed): ~$122B raise closes at $852B valuation (SoftBank, Microsoft, Nvidia participating) (Tech Insider). - 2026-04-08 (confirmed): CFO Friar states OpenAI will allocate IPO shares to retail investors (CNBC). - 2026-05 (reported): Friar reportedly pushing internally to delay IPO to 2027 over spending/reporting readiness (Forbes). - 2026-06-08 (confirmed): Confidential draft S-1 filed with SEC (StartupHub.ai, CNBC). - 2026-06-26/28 (reported): Multiple outlets report OpenAI "pushing IPO to 2027" amid market uncertainty (Yahoo, Breitbart). - 2026-07-08 (confirmed): BofA extends $520M loan to OpenAI ahead of IPO (Reuters via multiple outlets). - 2026-07-12 (reported): Fool.com reports Altman "waiting for $1T valuation"; notes SoftBank $40B loan due March 2027 (a soft external deadline pressure). - 2026-07-29/31 (confirmed): Friar tells employees July ARR topped all of Q2; enterprise revenue now bigger than consumer (CNBC). - 2026-08-11 (confirmed): OpenAI completes ~$7B employee share buyback at $852B valuation (IBTimes, StartupHub.ai) — confirms no new valuation mark since March 2026. - 2026-08-14 (confirmed): Executive departures (incl. revenue chief Denise Dresser) reported as "red flag" ahead of IPO (CNBC, Epoch Times, 247WallSt). - 2026 (ongoing, reported): Rival Anthropic's secondary-market valuation reportedly reaches $1.2T (Jul 2026) then $1.5T (Aug 2026) (Business Insider) — comparable escalating further ahead of its own possible 2026 IPO. # Event Will OpenAI's market cap close above $1T on its first day of public trading (resolves No if no IPO by Dec 31, 2027)? # Outcomes to forecast Yes / No # Kalshi market anchor No kalshi_direct price was returned in this research pass (tool output absent for this specific ticker). Nearest available direct read is the **Polymarket price for the identical question: 76% YES**, up from ~72.5% a month ago (30d +4.5pt) and ~72.5% a week ago (7d +3.5pt); 90-day range 62.5%–83.5%; volume ~$1.05M. This should be treated as the primary cross-market proxy anchor pending direct Kalshi confirmation. # Sub-question answers 1. **Polymarket price/trend** — 76% YES currently; steadily rising over 90 days (low 62.5%, high 83.5%), +4.5pt over 30 days (Polymarket direct). 2. **P(IPO completes by Dec 31 2027)** — No explicit probability found, but qualitative signals lean positive-but-uncertain: S-1 filed June 2026, banks (Goldman, Morgan Stanley, JPMorgan) engaged, but CFO favors 2027 over Altman's 2026 preference, and Aug 2026 exec turmoil raises delay risk (CNBC, StartupHub.ai, Forbes). 3. **Valuation figures** — Latest private mark: $852B (Aug 2026 buyback/secondary, unchanged since March 2026 raise). IPO target explicitly stated at ~$1T by Altman/reporting (NYT via Yahoo, TipRanks, Tech Insider). 4. **Restructuring/prerequisites** — PBC conversion and Microsoft agreement completed Oct 2025 (Wikipedia, Yahoo). Confidential S-1 filed June 2026; bankers hired; no confirmed CFO/auditor gap noted, though Friar flags readiness concerns (multiple sources). 5. **Adjacent markets implying conditional P(>$1T | IPO)** — "OpenAI IPO by Aug 31 2026" Polymarket market at 0.25% YES (near-certain no near-term IPO), consistent with 2027 timeline lean. Kalshi "OpenAI vs Anthropic IPO first" gives Anthropic 93% — suggesting market believes Anthropic likely lists before OpenAI, adding some timing risk to OpenAI's own 2027 deadline. 6. **Base rates / AI capex sentiment** — No explicit mega-cap IPO base-rate data returned; contextually, comparable AI peer Anthropic's secondary valuation surged from ~$380B to $1.2T–$1.5T within months (Jul–Aug 2026), suggesting strong sector tailwinds that could support a >$1T OpenAI IPO if listed in this environment (Business Insider, MarketWise). # Key facts (high-confidence, factual) 1. [Wikipedia/Yahoo] PBC conversion + Microsoft deal completed Oct 2025; valued OpenAI ~$500B then. 2. [StartupHub.ai/CNBC] Confidential S-1 filed June 8, 2026; no ticker/date set as of Aug 2026. 3. [IBTimes/StartupHub.ai] $852B valuation confirmed via Aug 2026 $7B employee buyback. 4. [Forbes/Startup Fortune/Yahoo] Internal Altman-Friar disagreement; company now leaning 2027 IPO. 5. [Tech Insider] 2025 net loss $38.5B on $13.07B revenue; cash burn ~$27B (2026e), ~$63B (2027e). # Cross-market signals - Kalshi related: Anthropic favored 93% to IPO before OpenAI (KXOAIANTH-40-ANTH), rising sharply (+10pt/7d) — timing competition signal, not valuation. - Polymarket: identical-question market at 76% YES; narrow "IPO by Aug 2026" market at 0.25% (confirms delay to 2027 consensus). - Sportsbook implied: N/A. # Analyst opinions and speculation - Altman: any IPO valuation <$1T is a "nonstarter" (Fool.com). - CFO Friar: internally cautious, prioritizing readiness over speed (multiple sources). - 247WallSt/CNBC: recent C-suite exits framed as a "huge red flag" for IPO prospects. - Analysts note Anthropic's rapid secondary-market re-rating to $1.2–1.5T could pull OpenAI's eventual IPO pricing upward via sector comps. # Directional lean per outcome - **Yes**: Explicit $1T floor target set by Altman; comparable AI valuations (Anthropic) already trading far above $1T; strong Polymarket YES (76%, rising); IPO prep (banks, S-1) well underway. - **No**: Current private mark ($852B) still below $1T threshold; CFO-driven delay risk could push IPO past 2027 deadline (auto-No); recent executive turmoil, large losses, and heavy cash burn add downside/timing risk; code-based bottom-up Fermi model (independent of market prices) estimated only ~19% joint probability — a notable divergence from the 76% market price, likely reflecting conservative assumptions on IPO-by-2027 probability and being an internal estimate rather than latest market info. # Gaps / unknowns - No direct Kalshi YES price returned for this exact ticker this pass. - No confirmed IPO date, exchange, or roadshow timeline. - Unclear how much premium/discount first-day pricing typically holds vs. last private mark for mega-cap tech IPOs (no base-rate data retrieved). - Large gap between market-implied ~76% and bottom-up Fermi ~19% unresolved; likely the market is pricing higher IPO-completion odds and higher conditional >$1T probability than the Fermi grid assumed. # Calibration anchors - Polymarket cross-market price (proxy anchor): 76% YES, trending up. - Fermi/base-rate model midpoint: ~19% (wide 90% CI 11–29%) — likely stale/conservative vs. current market consensus. - Precedent: Anthropic secondary valuations near-doubled to tripled within months (Jul→Aug 2026), illustrating how fast AI valuations can escalate pre-IPO.

1. Both forecasts converge on identical point estimates (0.70/0.30) despite claiming independent Fermi-style decomposition, suggesting anchoring on the Polymarket 76% figure rather than genuinely independent reasoning — neither explains why their conditional multiplication (e.g., 0.80×0.85≈0.68 vs 0.85×0.85≈0.72) should land at exactly the same headline number as the other. 2. Neither forecaster grapples seriously with the ~19% Fermi/base-rate estimate beyond dismissing it as "stale" — given the brief flags this as a genuine unresolved divergence, a more rigorous treatment would decompose *why* the bottom-up model differs (e.g., does it weight the SoftBank $40B March-2027 loan as forcing a lower/faster IPO, or model higher delay-past-2027 risk?) rather than simply discounting it. 3. Both assign a very high (~85%) conditional P(>$1T | IPO) largely on the strength of Altman's "nonstarter" rhetoric and Anthropic comps, but this ignores the SoftBank loan due March 2027 as a potential *forcing function* that could push OpenAI to IPO below $1T out of liquidity necessity rather than wait for a favorable valuation — a scenario neither model considers. 4. Neither forecast addresses IPO-pricing mechanics explicitly: first-day close typically reflects a "pop" over the offer price, so the relevant threshold isn't just "will they price at $1T" but "will trading close above it" — this could argue for a *higher* conditional probability than either forecaster's ~85% if underwriters price conservatively below the $852B–$1T gap and a pop pushes it over. 5. Both underweight the recent negative-momentum signals (Denise Dresser departure and other "red flag" exec churn reported Aug 14, 2026, CFO-Altman public disagreement) as evidence that IPO timing could slip past Dec 31, 2027 entirely (auto-No), rather than just causing delay within 2027 — the ~80-85% P(IPO by 2027) estimate may be too high given CFO's explicit preference for even more delay and no confirmed date/exchange/roadshow as of the brief. 6. Neither forecaster discusses the Anthropic-first timing race (93% Kalshi) as a signal of sector-wide IPO market saturation/absorption risk — if Anthropic IPOs first at $1.2-1.5T, this could either help (comp support) or hurt (investor fatigue, capital rotation) OpenAI's subsequent pricing, and this asymmetry isn't modeled.

70% (52% confidence)

Yes: 70% No: 30%

The resolution is a joint event: an IPO must complete by Dec 31, 2027 AND close day one above $1T. Preparation is genuinely advanced (confidential S-1 June 2026, Goldman/MS/JPM engaged, BofA bridge loan, retail allocation plan, SoftBank $40B loan due March 2027 creating funding pressure), supporting ~75-82% for a listing by end-2027, though CFO Friar's explicit delay preference, exec churn, and the absence of any date/exchange/roadshow keep meaningful slippage-past-2027 risk. Conditional on listing, >$1T is likely (~85-90%): Altman treats sub-$1T as a nonstarter, the $852B mark is stale since March 2026, Anthropic comps sit at $1.2-1.5T, and first-day pops mean the closing print can clear $1T even on a conservative offer price — the critique's pricing-mechanics point pushes this conditional up, partially offsetting the liquidity-forced-cheap-IPO scenario. That multiplies to ~0.66-0.72, so I settle at 0.70, modestly below the 76% Polymarket proxy (which may under-price the auto-No delay branch) and far above the stale ~19% bottom-up Fermi, whose main divergence appears to be an overly pessimistic IPO-completion term now contradicted by the actual S-1 filing.

Total pipeline time: 204.1s