Tools chosen: polymarket_direct polymarket_related kalshi_related claude_news claude_news gdelt_news wikipedia code_execution

Sub-questions (Fermi decomposition)

Planner reasoning

This is a Polymarket-sourced relative-valuation question resolving off Nasdaq Private Market (NPM) secondary-market prices for Anthropic vs OpenAI on Dec 31, 2026. The dominant drivers are the current NPM-implied valuations, the most recent primary/secondary funding rounds for each company, and revenue growth trajectories; OpenAI has historically carried a much larger valuation (~$500B+ vs Anthropic's ~$183-350B), so the base case strongly favors OpenAI unless the market's Yes/No mapping says otherwise. I need the direct market price first to anchor and to determine which outcome maps to which company.

# Current state As of mid-August 2026, Anthropic's last confirmed private valuation ($965B, May 2026 Series H) and secondary-market trading levels (reportedly $1.1T–$1.5T through Aug 2026) exceed OpenAI's last confirmed valuation ($852B, set by the March 2026 $122B round and reaffirmed by an August 2026 $7B employee buyback at the same mark). Polymarket's "Yes" (Anthropic higher) trades at 93%. Resolution depends specifically on NPM-published prices (or post-IPO market cap) as of Dec 31, 2026 — not on the current lead, which could still flip via a new OpenAI round, an IPO repricing, or continued Anthropic outperformance. # Timeline of key events - 2025-09: Anthropic raises at $183B valuation (reported). - 2025-10: OpenAI $6.6B share sale values company at $500B (confirmed, Wikipedia/press). - 2026-02: Anthropic raises $30B Series G at $380B valuation (reported). - 2026-03-31: OpenAI closes $122B round at $852B valuation, led by SoftBank/Amazon/Nvidia (confirmed, CNBC). - 2026-05-28: Anthropic closes $65B Series H at $965B post-money, overtaking OpenAI as most valuable AI startup (confirmed, multiple sources). - 2026-06-08: OpenAI confidentially files for U.S. IPO, targeting >$1T valuation (reported, Reuters). - 2026-06 (undated): Anthropic also confidentially files IPO prospectus; both eyeing Fall/Winter 2026 listings (reported, Forbes). - 2026-07-09: Anthropic secondary-market valuation reported at $1.2T (reported, Business Insider). - 2026-08-11: OpenAI completes $7B employee buyback, reaffirming $852B valuation — no new higher round since March (confirmed, IBTimes/Bloomberg via claude_news). - 2026-08-14: Anthropic secondary-market valuation reported at $1.5T; Anthropic reportedly eyeing $2T IPO (reported, Business Insider/Fortune). # Event Which company — Anthropic or OpenAI — has the higher NPM-reported (or public market cap, if IPO'd) valuation on Dec 31, 2026. # Outcomes to forecast Yes (Anthropic higher) / No (OpenAI higher) # Kalshi market anchor No kalshi_direct price was returned in research; only Polymarket data available for this exact market. **Polymarket YES (Anthropic higher) = 93%**, up +5.5% (7d) and +11.5% (30d), range 63.5%–94.5% over 89 days, volume $52,488. Trend has been consistently rising toward Anthropic. Related Kalshi market "Will OpenAI or Anthropic IPO first? — Anthropic" also prices at 93%, up sharply from 22% low — directionally consistent. # Sub-question answers 1. **Polymarket price / Yes mapping** — Yes = Anthropic higher; currently 93%, rising over past 30 days. [polymarket_direct] 2. **Latest NPM/private valuations & gap** — Anthropic: $965B (May 2026 Series H, confirmed) with secondary trades reported up to $1.1T–$1.5T (Jul–Aug 2026, reported). OpenAI: $852B (March 2026 round, reaffirmed Aug 2026 buyback, confirmed). Gap: Anthropic leads by ~$113B on confirmed rounds, or up to ~$650B on secondary marks. [claude_news, valueaddvc.com, businessinsider.com] 3. **Latest funding rounds** — OpenAI: $122B raise closing 3/31/2026 at $852B (confirmed); earlier $500B mark (Oct 2025). Anthropic: $183B (Sep 2025) → $380B (Feb 2026) → $965B (May 2026, confirmed); reportedly in talks for $900B–$2T range subsequently. [claude_news, Wikipedia] 4. **Revenue run-rates** — Anthropic: officially $47B run-rate (May 2026); unofficial estimates (Benioff) put it at $74.1B by Aug 2026. OpenAI: ~$2B/month (~$24-25B annualized) per company statements; one estimate cites $40-41B run-rate (Aug 2026, unconfirmed). Anthropic appears to be closing/exceeding the gap on growth. [claude_news, 247wallst.com] 5. **IPO timing** — Both have confidentially filed for IPOs (OpenAI 6/8/2026; Anthropic mid-2026), both targeting Fall/Winter 2026 listings; OpenAI reportedly considering delay to 2027. A forecasting model (futuresearch.ai) has Anthropic IPOing first (median Nov 30, 2026, ~$1.18T first-day cap) with OpenAI following mid-2027 (~$1.0T). If this holds, Anthropic would be public and priced above OpenAI's private mark by Dec 31, 2026. [claude_news, futuresearch.ai] 6. **Swing needed / base rates** — Code-execution Monte Carlo (using apparently stale/hypothetical inputs, Anthropic $300B vs OpenAI $650B) implies only 5-20% crossing probability under neutral drift — but this conflicts with confirmed data showing Anthropic already ahead as of Aug 2026. This tool's base assumptions appear outdated; deprioritized in favor of confirmed valuation facts. [code_execution — flagged as stale] 7. **Structural risks** — OpenAI completed a restructuring (2025) reducing Microsoft dependence; both firms carry heavy capex/compute commitments. No confirmed down-round for either firm; Anthropic has diversified backers (Google, Amazon, Nvidia, Microsoft) vs OpenAI's core Microsoft/SoftBank/Nvidia backing. AI-bubble commentary exists but no confirmed devaluation event. [claude_news, Wikipedia] # Key facts (high-confidence, factual) 1. [CNBC/confirmed] OpenAI's last priced round: $852B (3/31/2026), reaffirmed by Aug 2026 buyback. 2. [Multiple, confirmed] Anthropic's last priced round: $965B (5/28/2026 Series H). 3. [businessinsider.com, reported] Anthropic secondary trades reached $1.2T (Jul 2026) and $1.5T (Aug 2026). 4. [Reuters/CNBC, reported] Both firms have confidentially filed for IPOs in 2026. 5. [Polymarket] Yes trading at 93%, rising trend. # Cross-market signals - Kalshi related: "OpenAI or Anthropic IPO first? — Anthropic" at 93%, up from 22% low — consistent directional signal. - Polymarket (this market): 93% Yes, +11.5% 30d. - No sportsbook signal available. # Analyst opinions and speculation - Headlines describe "Anthropic is compounding, OpenAI is flatlining" (Yahoo/Proactive, Aug 2026) — reflects momentum narrative, not confirmed valuation event. - Anthropic targeting $2T IPO valuation (Fortune, reported/speculative). - Futuresearch.ai model projects Anthropic IPOs first at higher implied cap than OpenAI's eventual listing. # Directional lean per outcome - **Yes (Anthropic higher)**: Supported by confirmed May 2026 round ($965B>$852B), secondary trades up to $1.5T, faster revenue growth, likely earlier/higher IPO pricing, and strong/rising Polymarket and Kalshi-adjacent pricing (93%). - **No (OpenAI higher)**: Supported only by scale/brand dominance, possible large new OpenAI round or IPO exceeding $1T+ before Dec 31, and inherent uncertainty in NPM data continuity/timing. # Gaps / unknowns - No direct kalshi_direct price captured for this specific ticker. - Exact NPM-reported (vs. secondary-market rumored) prices for each company are not fully disclosed in research. - IPO timing for both companies remains fluid; a public listing before Dec 31 would shift resolution to market cap, adding volatility. - Code_execution Monte Carlo used outdated/inconsistent inputs — should not be weighted heavily. # Calibration anchors - Polymarket Yes price: 93% (rising trend, high-confidence anchor). - Precedent: Anthropic already overtook OpenAI in a prior valuation race (May 2026) after starting far behind — shows large multi-month swings are possible in this sector.

1. Both forecasters cluster tightly at 91–92%, essentially anchoring on Polymarket's 93% without independently stress-testing the largest identified risk: OpenAI's own confidential IPO filing (6/8/2026) explicitly targets ">$1T valuation" — if OpenAI IPOs before Anthropic or at a materially higher first-day cap, this alone flips resolution regardless of Anthropic's current private lead, and neither rationale quantifies this scenario distinctly from "a new OpenAI round." 2. Neither forecast addresses the resolution-mechanics risk flagged in the brief: NPM-reported prices vs. secondary-market "reported" trades ($1.1T-$1.5T) are not confirmed valuations, and if the official Dec 31 NPM/round-based comparison reverts to confirmed marks ($965B vs $852B) while OpenAI closes even a modest new round (e.g., $1T+), the gap could close faster than the secondary-market narrative suggests — this ambiguity in what counts as "the" valuation on resolution day deserves more weight than a rounding-error adjustment. 3. Both treat the futuresearch.ai model (Anthropic IPOs first, ~$1.18T cap, OpenAI not until mid-2027 at ~$1.0T) as corroborating evidence, but if OpenAI's IPO timeline actually slips to 2027 as noted ("OpenAI reportedly considering delay to 2027"), that removes the main repricing catalyst for OpenAI before Dec 31 — this is actually a bullish structural factor for Yes that could justify pushing above 93%, yet both forecasters shaded *below* the anchor instead of exploring upside. 4. The brief's Monte Carlo tool implied only 5-20% crossing probability, explicitly flagged as stale — but neither forecaster explains why they're comfortable dismissing it entirely rather than assigning it even minimal residual weight as a hedge against unknown-unknowns in valuation methodology. 5. Neither rationale discusses potential AI-sector-wide shock (bubble correction, capex/compute cost blowups) that could hit both firms simultaneously but asymmetrically — since Anthropic's premium is more secondary-market/momentum-driven (less "hardened" by confirmed primary rounds) versus OpenAI's, a sector selloff could disproportionately compress Anthropic's speculative secondary marks back toward its $965B primary print, narrowing or reversing the gap. 6. Both converge within 1pp of each other and within 1-2pp of the Polymarket anchor, showing insufficient independent variance — given the explicit uncertainty about IPO timing volatility ("adding volatility") and unconfirmed secondary pricing, a wider spread of scenario-weighted outcomes (e.g., 85-95% range across forecasters) would better reflect the genuine tail risks rather than both essentially restating the crowd price with minor rhetorical hedging.

92% (71% confidence)

Yes: 92% No: 8%

Anthropic already leads on confirmed primary marks ($965B May 2026 vs OpenAI's $852B, merely reaffirmed by the August 2026 buyback), and secondary trades reportedly $1.1T–$1.5T imply NPM-published prices well above OpenAI's. For 'No' to resolve, OpenAI must both reprice materially upward (new round or IPO above $1T) AND have that reflected by Dec 31 — plausible given its confidential filing, but reports of a possible slip to 2027 reduce the likelihood of a pre-deadline catalyst. The devil's advocate correctly notes the OpenAI-IPO and sector-shock paths (which could compress Anthropic's momentum-driven secondary premium back toward its primary print) but these are largely offset by the upside case that OpenAI has no repricing event at all before year-end; the stale Monte Carlo carries little weight since its inputs contradict confirmed facts. I therefore sit essentially at the 93% Polymarket anchor, shading a touch for ~4.5 months of IPO-timing volatility and resolution-source ambiguity.

Total pipeline time: 243.2s