Tools chosen: polymarket_direct claude_news gdelt_news polymarket_related kalshi_related wikipedia code_execution

Sub-questions (Fermi decomposition)

Planner reasoning

This is a Polymarket head-to-head market where the base case is likely neither company IPOs before end-2027 (resolving 50-50), so the key research is the relative probability and timing of each company's IPO preparations. I need the current market price as anchor, then recent reporting on OpenAI's restructuring/IPO groundwork (Reuters reported ~$1T IPO prep as early as H2 2026) and Anthropic's own IPO signals (CFO hires, Wilkinson/advisers, funding rounds). Cross-venue and related markets give triangulation on IPO timing probabilities.

# Current state Both OpenAI and Anthropic have filed confidential draft S-1s (Anthropic June 1, 2026; OpenAI June 8, 2026), but neither has priced or completed an actual IPO. Anthropic has since moved further along (investor meetings scheduled mid-July 2026, credit-line arrangements), while OpenAI's timeline is contested internally (CFO Friar wants 2027, CEO Altman wants faster) and reportedly wobbling amid a "SpaceX scare" delay report. No IPO has closed for either company as of the latest research (Aug 2026). # Timeline of key events - 2026-05-15/21: Early reports OpenAI "to file IPO paperwork in coming days"; hiring signals at Anthropic for PE/finance roles (reported). - 2026-06-01: Anthropic confidentially files draft S-1 with SEC at ~$965B valuation (confirmed, multiple outlets). - 2026-06-08: OpenAI confidentially files draft S-1; company statement says timing undecided (confirmed, OpenAI official post). - 2026-06-08: Wedbush note frames Anthropic's filing as a move to "get ahead" of OpenAI (reported analyst take). - Mid-June 2026: Reports OpenAI hired Goldman Sachs & Morgan Stanley, targeting ~Sept 2026 debut at $730–850B valuation (reported, uncertain per OpenAI's own caveats). - 2026-06-26/27: Reports of possible OpenAI IPO delay following "SpaceX scare"; SoftBank shares dip on delay report (reported). - 2026-07-15/17: Anthropic reportedly scheduling investor meetings, listing seen as nearing (reported, multiple outlets). - 2026-07-16/21: Anthropic reportedly pursuing multibillion-dollar pre-IPO credit lines, "SpaceX playbook" (reported). - 2026-08-06/09: Anthropic valuation figures ($965B–$1T range) circulate in financial press; OpenAI's public S-1 still not on EDGAR as of Aug 9, 2026 (confirmed absence); CFO Friar reportedly advising delay to 2027 (reported, internal conflict with Altman). # Event Will Anthropic or OpenAI complete an IPO first (by Dec 31, 2027)? Ticker outcomes are Yes/No, mirroring "Anthropic first" (Yes) vs "OpenAI first / other" (No); 50-50 split applies if neither IPOs, both IPO same day, or both become unable to IPO. # Outcomes to forecast - Yes (Anthropic IPOs first) - No (OpenAI IPOs first, or 50-50/void conditions favor No pricing) # Kalshi market anchor No direct kalshi_direct pull returned for this exact ticker; closest same-topic Kalshi market (KXOAIANTH-40-ANTH, "Anthropic" side) trades at **92% YES**, up +10pts over both 7d and 30d, avg volume ~1,158 contracts/day, 90-day range 22%–92% — a large recent run-up. This should be treated as the primary anchor given ticker equivalence, though note it may not be the identical contract. # Sub-question answers 1. **Polymarket price/trend** — Polymarket's equivalent market prices "Anthropic first" at 84.5%, down -5% over 7 days, roughly flat over 30 days; 90-day range was 24.5%–91%, implying a huge swing from OpenAI-favored to Anthropic-favored pricing at some point (Polymarket direct). 2. **OpenAI IPO prep** — Confidential S-1 filed June 8, 2026; hired Goldman Sachs & Morgan Stanley; rumored Sept 2026 target at $730–850B valuation; but OpenAI's own statement says timing undecided, and CFO Sarah Friar reportedly pushing internally for a 2027 delay against Altman's faster preference (claude_news, gdelt_news). 3. **Anthropic IPO prep** — Confidential S-1 filed June 1, 2026 (one week ahead of OpenAI) at $965B valuation; CFO Krishna Rao (ex-Airbnb IPO lead) steering process; investor meetings scheduled mid-July 2026; pursuing multibillion-dollar pre-IPO credit lines "SpaceX playbook" (claude_news, gdelt_news). 4. **Capital pressure** — Both firms hold enormous private valuations ($965B Anthropic, $500-850B OpenAI) and have been raising via private credit/equity rather than being forced to list; no explicit cash-crunch evidence found forcing near-term IPO for either; research is silent on hard capital constraints. 5. **Base rate for either IPOing by 2027** — Not directly quantified by news; code_execution model estimates plausible P(OpenAI IPO by 2027)≈15-45%, P(Anthropic)≈3-15%, yielding P(neither)≈53-82% in most scenarios — suggesting the "50-50" clause carries substantial weight. 6. **Acquisition/merger triggering 50-50** — No evidence found of Microsoft/Amazon/Google absorption plans; both remain independent PBCs pursuing IPOs, not acquisition targets (wikipedia, gdelt_news). # Key facts (high-confidence, factual) 1. [claude_news/openai.com] OpenAI confidentially filed draft S-1 June 8, 2026; no committed listing date. 2. [claude_news/useluminix] Anthropic confidentially filed draft S-1 June 1, 2026, one week before OpenAI, at ~$965B valuation. 3. [techjournal.org] OpenAI's public S-1 had not appeared on EDGAR as of Aug 9, 2026. 4. [wikipedia] Anthropic valued ~$965B (May 2026); OpenAI valued ~$500B after Oct 2025 share sale, with later banker targets of $730-850B. 5. [gdelt_news] Anthropic reported to be scheduling investor meetings and arranging pre-IPO credit lines (July 2026), signaling advanced process stage. # Cross-market signals - Kalshi related (KXOAIANTH-40-ANTH): 92% Anthropic-first, +10pts/7d and /30d — strong recent momentum toward Anthropic. - Polymarket: 84.5% Anthropic-first, -5%/7d, roughly flat/30d — also favors Anthropic but less extreme and slightly cooling. - No sportsbook data available. - Divergence (92% vs 84.5%) suggests some cross-venue disagreement or lag; Kalshi shows more bullish/less liquid pricing (lower volume than headline events). # Analyst opinions and speculation - Wedbush: Anthropic's filing is a competitive "get ahead of OpenAI" move. - Bankers (unnamed, biggo.com): frame this as winner-take-all race where first-to-market "defines the industry." - Internal OpenAI conflict reported: CFO Friar wants 2027 delay citing reporting-readiness concerns; Altman wants fastest possible timeline — unresolved as of Aug 2026. # Directional lean per outcome - **Yes (Anthropic first)**: Filed S-1 one week earlier, more concrete recent steps (investor meetings, credit lines), both Kalshi and Polymarket price it as strong favorite (84.5-92%). - **No (OpenAI first / other)**: OpenAI has bigger banks (Goldman/Morgan Stanley), higher target valuation range, but faces internal timeline conflict and a reported delay scare; public S-1 still pending as of Aug 2026, working against near-term OpenAI lead. # Gaps / unknowns - No confirmed actual IPO date for either company; all reporting reflects private/confidential process stages only. - Uncertain whether 50-50 clause (neither IPOs by 2027) is being fully priced into the Yes/No binary structure of this specific ticker — model estimates P(neither)=53-82%, which is large. - Discrepancy between Kalshi (92%) and Polymarket (84.5%) unresolved; could reflect different underlying contract structures or timing of quotes. # Calibration anchors - Kalshi-equivalent YES (Anthropic) price: 92% (recent +10pt momentum). - Polymarket YES (Anthropic) price: 84.5% (recent -5% pullback). - Model-based fair value range for Anthropic-first: 0.29–0.50 (central case ~0.39), based on competing-hazard IPO-timing simulation — notably below both market prices, suggesting markets may be overpricing Anthropic or underweighting the "neither" scenario.

1. Both forecasts land near the market anchors (0.85, 0.80) but barely engage with the brief's own model-based fair value range of 0.29–0.50 (central ~0.39) for Anthropic-first — a huge gap they wave away in one sentence without adjusting much; if the model correctly captures P(neither)≈53-82%, the 50-50 split should pull Yes toward ~0.5-0.6, not 0.80-0.85. 2. Neither forecaster explicitly walks through the arithmetic of how the 50-50 clause interacts with P(neither)=53-82%: if P(neither) is that large, even a strong Anthropic lead among "someone IPOs" scenarios yields Yes≈P(Anthropic first)+0.5*P(neither), which could plausibly be far lower than 0.80 if P(Anthropic first | someone IPOs) isn't overwhelming. 3. Both treat Kalshi (92%) and Polymarket (84.5%) as converging confirmation, but the brief flags unresolved divergence and low Kalshi volume (~1,158 contracts/day) plus explicit uncertainty about whether this is even the identical contract — this warrants more skepticism/wider discounting than a simple "split the difference" move. 4. Neither forecast addresses the "SpaceX scare" delay as potentially informationally stale or overstated — it's reported only via secondhand rumor, yet both use it as a load-bearing reason for OpenAI lagging, without weighing that Anthropic's own credit-line/investor-meeting signals are similarly rumor-based ("reported," not confirmed). 5. Both forecasts anchor confidence too tightly to market prices without independently re-deriving a base rate check: the brief's own sub-question 5 estimates are wide (P(OpenAI by 2027)≈15-45%, P(Anthropic)≈3-15%) — notably Anthropic's estimated probability of IPOing at all is lower than OpenAI's in that model, which contradicts the "Anthropic heavy favorite" framing both forecasters adopt. 6. Tail scenario underweighted: neither addresses possibility of regulatory/SEC delay, market conditions (e.g., 2027 downturn), or one company being acquired/restructured (PBC conversion litigation) before Dec 2027, which could push toward "neither" more than the discussed base rates suggest.

79% (56% confidence)

Yes: 79% No: 21%

Anthropic is demonstrably further along the same path: S-1 filed a week earlier, investor meetings scheduled, pre-IPO credit lines being arranged, and an experienced IPO CFO, while OpenAI's public S-1 remains absent from EDGAR amid a reported internal Altman/Friar timing conflict. Both venues price Anthropic-first heavily (Kalshi 92%, Polymarket 84.5%), and I weight that real-money consensus above the brief's simulation, whose inputs (P(Anthropic IPO by 2027)≈3-15%) look badly stale given a filed S-1 and active roadshow prep. The red team is right, however, that the 50-50 'neither IPOs' clause is a genuine drag: with even a 20-30% chance no listing prices by end-2027, Yes decomposes to P(Anthropic first)+0.5*P(neither), which caps realistic values well below the 92% Kalshi tick unless Anthropic's conditional lead is near-total. Additional discounts for possible contract non-identity, thin Kalshi volume, rumor-based (not confirmed) reporting on both sides, and tail risks (market downturn, PBC/restructuring litigation, SEC delay) justify sitting slightly below Polymarket rather than at the Kalshi anchor.

Total pipeline time: 166.8s