Tools chosen: polymarket_direct claude_news claude_news polymarket_related kalshi_related gdelt_news wikipedia

Sub-questions (Fermi decomposition)

Planner reasoning

This is a Polymarket question comparing Anthropic's vs OpenAI's private valuations on NPM as of Dec 31, 2026. The key drivers are current NPM secondary-market pricing, recent/rumored primary funding rounds (OpenAI ~$500B, Anthropic ~$350B), and any IPO plans. I need the market anchor, current news on both companies' valuations, and cross-venue signals.

# Current state As of late August 2026, Anthropic's NPM/secondary-market valuation (~$1.1–1.5T per multiple secondary sources) exceeds OpenAI's latest primary-round valuation (~$852B, reconfirmed via August tender). Resolution depends on NPM Price data (or public market cap if either IPOs) specifically as of Dec 31, 2026 — over a year away from the question's actual close date implied by data (period ends 2026-12-31; today's research is dated Aug 2026), so significant time remains for both trajectories, new rounds, and possible IPOs to shift the outcome. # Timeline of key events - 2026-01-07: Anthropic signs term sheet for $10B round at $350B valuation (confirmed, cnbc.com). - 2026-03-31: OpenAI closes $122B round at $852B post-money valuation (confirmed, valueaddvc.com). - 2026-05-28: Anthropic closes Series H, $65B raised at $965B post-money — first time Anthropic's primary valuation exceeds OpenAI's (confirmed, anthropic.com). - 2026-06-01: Anthropic confidentially files S-1 with SEC (confirmed, futuresearch.ai). - 2026-06-08: OpenAI confidentially files S-1 (confirmed, futuresearch.ai). - 2026-06-25/28: Reports OpenAI may delay IPO to 2027 after SpaceX's rocky debut (reported, Forbes/Benzinga). - 2026-07-09/10: Secondary markets price Anthropic at ~$1.2T, "overtaking OpenAI" (reported, Business Insider/TheNextWeb). - 2026-07-20: OpenAI sees "resurgence of interest" in secondary markets (reported, Business Insider). - 2026-08-10/11: OpenAI completes $7B employee tender offer, reconfirming $852B valuation (confirmed, TechCrunch/ibtimes). - 2026-08-12: NPM prices Anthropic at $702.07/share (~$1.1–1.2T implied) (confirmed, nasdaqprivatemarket.com). - 2026-08-13/14: Reports Anthropic targeting ~$2T IPO valuation in October 2026, potentially largest IPO ever (reported, Fortune/Forbes); Business Insider cites $1.5T secondary valuation same week (reported — note internal inconsistency across sources, $1.1T–$1.5T range). - 2026-08-20: OpenAI CFO Sarah Friar tells staff company targets public listing "by 2027" (reported, sacra.com). - 2026-08-25: Forge Global prices OpenAI secondary shares at $721.85 (~consistent with $850B range) (confirmed, forgeglobal.com). - Ongoing: Kalshi market "Will OpenAI or Anthropic IPO first?" prices Anthropic at 94% (up from 24% 90 days ago), suggesting strong market consensus Anthropic lists first. # Event Kalshi/Polymarket cross-listed event: which company (Anthropic or OpenAI) has the higher valuation, per NPM Price (or public market cap post-IPO), as of Dec 31, 2026. # Outcomes to forecast - Yes (Anthropic higher valuation) - No (OpenAI higher valuation) # Kalshi market anchor No direct Kalshi YES price was returned for this specific ticker in raw research (kalshi_direct tool output absent); Polymarket equivalent shows current price 90% (Yes/Anthropic), down 2.5% over 7 days but up 6.5% over 30 days, range 63.5%–94.5%, volume ~$57K (thin). Treat 90% as best available cross-venue anchor given no direct Kalshi print was retrieved. # Sub-question answers 1. **Current NPM prices/gap** — Anthropic ~$1.1–1.5T (NPM: $702.07/share ~Aug 12; secondary reports range $1.1T–$1.5T) vs OpenAI ~$852B (Forge: $721.85/share, Aug 25, consistent with tender price). Gap: Anthropic ahead by roughly $250B–$650B depending on source (valueaddvc.com, nasdaqprivatemarket.com, forgeglobal.com, Business Insider). 2. **Latest primary rounds** — OpenAI: $122B round closed 2026-03-31 at $852B, reconfirmed via $7B tender Aug 2026 at same price. Anthropic: Series H closed 2026-05-28, $65B raised at $965B post-money — no update indicating a newer primary round since then (anthropic.com, valueaddvc.com). 3. **IPO plans before Dec 31, 2026** — Both filed confidential S-1s (Anthropic June 1, OpenAI June 8, 2026). Anthropic reportedly targeting Sept/Oct 2026 listing at $1.82T–$2T valuation (Fortune, KuCoin, futuresearch.ai forecast median Oct 26, 2026). OpenAI reportedly delaying to 2027 per Reuters/CFO comments (Forbes, sacra.com). If Anthropic IPOs in 2026, resolution would shift to public market cap, likely locking in a large Anthropic lead. 4. **6-12 month valuation trend** — Anthropic has moved from $350B (Jan 2026) → $965B (May) → $1.1–1.5T (secondary, Aug), a rapid ascent; OpenAI static at $852B since March 2026 (three sources: reconfirmed via tender in August at same price). Gap is widening in Anthropic's favor, not closing. 5. **Revenue run-rates** — Anthropic: $47B ARR (mid-May) → $65B annualized run-rate (end of July 2026); bankers reportedly underwriting to $190-200B 2028 revenue (futuresearch.ai, CNBC). OpenAI: ~$13B 2025 revenue, ~$20B+ annualized recently, $20.9B operating loss (valueaddvc.com, Reuters). 6. **Cross-venue signals** — Polymarket prices this exact event at 90% Yes (Anthropic). Related Kalshi market "Will OpenAI or Anthropic IPO first?" shows Anthropic at 94% (up sharply from 24% three months ago), reinforcing expectation Anthropic both IPOs first and likely retains higher valuation. # Key facts (high-confidence, factual) 1. [anthropic.com] Anthropic Series H, $65B raised, $965B post-money, closed 2026-05-28. 2. [valueaddvc.com] OpenAI $122B round, $852B post-money, closed 2026-03-31; reconfirmed via $7B tender Aug 2026. 3. [nasdaqprivatemarket.com] NPM Anthropic share price $702.07 as of 2026-08-12. 4. [forgeglobal.com] Forge OpenAI share price $721.85 as of 2026-08-25. 5. [futuresearch.ai] Both companies confidentially filed S-1s in June 2026. 6. [Kalshi related market] "OpenAI or Anthropic IPO first — Anthropic" priced 94%, up from 24% 90 days prior. # Cross-market signals - Kalshi related: IPO-first market strongly favors Anthropic (94%), consistent with Anthropic maintaining valuation lead if it IPOs first at a premium. - Polymarket (this market): 90% Yes (Anthropic), volatile 30-day range (63.5%-94.5%), thin volume ($57K) — low liquidity, moderate confidence signal. - No direct sportsbook signal available; no additional Polymarket-related matches found. # Analyst opinions and speculation - Fortune/KuCoin: Anthropic IPO could value it at $1.82T–$2T, "largest IPO ever," eclipsing SpaceX — speculative, unconfirmed pricing. - Forbes: questions whether $2T Anthropic valuation signals "AI bubble territory." - futuresearch.ai forecasting model: median Anthropic IPO Oct 26, 2026 at $1.82T; OpenAI IPO median July 2027 at ~$1.06T — if accurate, OpenAI wouldn't even be public by Dec 31, 2026, keeping NPM as resolution source with Anthropic far ahead. - Some reports (Business Insider, Aug 14) cite $1.5T Anthropic secondary valuation vs. NPM's implied ~$1.1-1.2T — source discrepancy suggests valuation marks are volatile/inconsistent across secondary platforms. # Directional lean per outcome - **Yes (Anthropic)**: Strong support — primary round lead since May 2026 ($965B vs $852B), secondary/NPM marks extending lead to $1.1-1.5T, 2x revenue run-rate, faster IPO timeline (potentially locking in premium before Dec 31), corroborated by both Polymarket (90%) and related Kalshi IPO-race market (94%). Opposing: OpenAI could announce a large new round or accelerate IPO; single-source valuation claims (e.g., $2T) unconfirmed; if OpenAI IPOs at outsized valuation before Anthropic, dynamics could flip, though current reporting suggests OpenAI more likely delayed to 2027. - **No (OpenAI)**: Weak support — static $852B valuation since March, revenue lagging, IPO reportedly slipping to 2027. Would require a major new mega-round, an accelerated/premium IPO, or Anthropic's IPO faltering/pricing below expectations. # Gaps / unknowns - No direct Kalshi YES price for this exact ticker was retrieved (only Polymarket 90% and a related Kalshi IPO-race market at 94%) — should confirm actual Kalshi print. - No confirmed data closer to the actual Dec 2026 resolution date; all evidence is from Jan–Aug 2026, leaving ~4 months of unknown developments. - Conflicting secondary valuation figures for Anthropic ($1.1T NPM vs $1.5T Business Insider vs $2T IPO speculation) not reconciled. - Uncertain whether either IPO actually completes before Dec 31, 2026, which would change resolution methodology to public market cap. # Calibration anchors - Polymarket current price: 90% Yes (Anthropic) — primary cross-venue anchor in absence of direct Kalshi data. - Related Kalshi market (IPO-first, Anthropic): 94%, up from 24% three months ago — directionally reinforces Anthropic lean. - Precedent: gap has only widened in Anthropic's favor over 8 months tracked (Jan $350B → May $965B → Aug $1.1-1.5T), while OpenAI has been flat at $852B since March — momentum favors continuation of Anthropic's lead through year-end absent a shock event.

1. Both forecasters converge suspiciously precisely on 0.910/0.090, essentially just nudging the 90% Polymarket anchor up slightly with near-identical reasoning—neither independently stress-tests the anchor itself, which is thin-volume ($57K) and has swung 63.5%–94.5% in 30 days, suggesting the "true" probability could be less stable than the point estimate implies. 2. Neither forecast adequately weights the ~4-month information gap before resolution (brief explicitly flags "no confirmed data closer to the actual Dec 2026 resolution date"); given OpenAI's history of large surprise rounds ($122B round, $7B tender) in short windows, a mega-round or valuation reset before Dec 31 is a live tail risk that could compress or flip the gap, warranting more than a passing mention. 3. Both treat the IPO-timing dynamic as uniformly favorable to Yes, but neither considers the scenario where Anthropic's IPO actually prices below the speculative $1.8–2T range (noted as "unconfirmed pricing" and flagged by Forbes as possible "bubble territory")—a disappointing IPO print could simultaneously become the resolution source AND undercut the lead, a scenario that cuts against Yes despite superficially being "Anthropic IPOs first." 4. The brief's internal valuation inconsistency ($1.1T NPM vs $1.5T Business Insider vs $2T IPO speculation) is acknowledged by both forecasters as noise, but neither considers whether this volatility itself increases variance around Dec 31 mark-to-market, which is asymmetric risk since OpenAor has been comparatively stable at $852B—stable reference points arguably deserve more weight in a point-in-time resolution than volatile ones. 5. Neither forecast explores the mechanical resolution-source risk: if only one company IPOs by Dec 31 (likely Anthropic per the brief) while the other remains NPM-priced, the comparison method changes qualitatively (public market cap vs. private secondary mark), which could introduce a mismatch not fully modeled by simply extrapolating current trends. 6. Both forecasts essentially anchor to Polymarket's 90% without independently deriving a fundamentals-based estimate first and then checking convergence—given the valuation gap is already $250B–$650B on primary rounds alone with widening revenue divergence ($65B vs $20B run-rate), a case could be made for underconfidence rather than the implied caution suggesting they're "not pushing higher" than the anchor.

92% (73% confidence)

Yes: 92% No: 8%

Anthropic leads on every measurable dimension: primary round ($965B vs $852B), secondary marks ($1.1–1.5T vs ~$852B), and revenue run-rate (~$65B vs ~$20B+), with the gap widening rather than narrowing across eight tracked months while OpenAI has been flat since March 2026. The critique's strongest points — the ~4-month information gap and the risk of a disappointing Anthropic IPO print — are real but largely insufficient to flip the outcome: even an IPO pricing well below the speculative $1.8–2T range (say $1.0–1.2T) would still exceed OpenAI's mark, so No essentially requires an OpenAI mega-round or premium listing that re-rates it above ~$1T, a live but minority scenario. Resolution-source mismatch (public cap vs NPM) adds noise but not clear directional risk given the size of the gap. The thin-volume Polymarket anchor at 90% is noisy, but an independently derived fundamentals view sits slightly above it, so I settle marginally higher at 0.92 rather than defaulting to the anchor.

Total pipeline time: 178.2s