Scan: Will the Hodge Conjecture be the next Millennium Prize Problem solved by AI?

1. Planner Fable 5.1 · Fermi decomposition + tool plan

Tools chosen: polymarket_direct polymarket_related kalshi_related article_search claude_news gdelt_news wikipedia

Sub-questions (Fermi decomposition)

What is the current Polymarket price for the Hodge Conjecture outcome and for the sibling outcomes (Riemann, P vs NP, BSD, Yang-Mills, no solution) in the same event?
Have any eligible AI labs (OpenAI, Google DeepMind, Anthropic, Harmonic, Axiom Math, Math Inc, etc.) publicly stated they are targeting Millennium Prize Problems, and which ones specifically?
What is the current state of AI mathematical research capability (e.g., Erdős problems, IMO gold, formal proof systems, AlphaProof/AlphaEvolve, GPT-5 math results) and how far is it from research-level open problems?
Is the Hodge Conjecture considered more or less tractable than the other four qualifying problems by mathematicians, and has there been any recent partial progress on it?
Have there been any premature or contested claims by AI labs of solving major open math problems (e.g., Erdős problem controversies), indicating the propensity for labs to announce a 'solution' even if unverified?
Does Kalshi or any other market have a related market on AI solving a Millennium Prize Problem, and does it agree with Polymarket?

Planner reasoning

This is a multi-outcome market where Hodge must be the FIRST Millennium Prize Problem announced as solved by an eligible AI lab before end of 2027. Key drivers are: whether any lab claims any solution at all by then (base rate near zero historically, but frontier math AI is advancing rapidly), and conditional on a claim, which problem is most tractable/likely to be targeted (Hodge is widely considered among the hardest and least amenable to computational or formal approaches compared to e.g. Riemann or BSD). The Polymarket price is the primary anchor; sibling markets on the other problems give the relative distribution.

2. Research Parallel tool calls

tool	status	time	items	summary
polymarket_direct	OK	0.3s	1	## This Market's Polymarket Data Will the Hodge Conjecture be the next Millennium Prize Problem solved by AI? - Current price (probability): 41.50% - 7-day price change: +15.00% - 30-day price change: +15.00% - Total volume: $32,024 (USD notional) - Price range: 26.50% - 42.00% - Data points: 4
polymarket_related	OK	2.0s	0	Scanned 100 active Polymarket markets, kept 0 matches. keyword 'Millennium Prize': 0 markets \| keyword 'Riemann Hypothesis': 0 markets \| keyword 'P vs NP': 0 markets \| keyword 'Hodge Conjecture': 0 markets \| keyword 'AI solve math': 0 markets
kalshi_related	OK	1.8s	0	0 related markets / summaries. keyword 'Millennium Prize': no matches \| keyword 'Riemann Hypothesis': no matches \| keyword 'AI math': no matches
article_search	OK	1.0s	29	Fetched 29 unique articles across 4 queries (lookback=120d). 'AI Millennium Prize Problem Hodge Conjecture': 10 articles \| 'AI lab solves open math problem Riemann Hypothesis': 10 articles \| 'OpenAI Google DeepMind mathematics research-level proof 2026': 10 articles \| 'Harmonic Axiom Math AI theorem
claude_news	OK	24.8s	12	## Findings: AI and the Hodge Conjecture / Millennium Prize Problems (as of Sept 2026) - OpenAI's claimed Navier-Stokes breakthrough (Sept 8, 2026): OpenAI announced in a blog post that an “internal OpenAI system” had found a solution to Navier-Stokes, using around 10,000 AI agents that worked
gdelt_news	OK	201.5s	10	GDELT: 10 articles across 3 queries (lookback=90d). 'Hodge conjecture AI': error GDELT rate-limited after retries (429) \| 'Millennium Prize Problem artificial intelligence': error GDELT rate-limited after retries (429) \| 'AI proves Riemann hypothesis': 10 hits
wikipedia	OK	0.1s	4	Fetched 4 Wikipedia entries (2 missing pages).

3. Evidence Brief Sonnet 5 · 8446 chars

# Current state As of Sept 2026, no eligible AI lab has made a qualifying (official, "solved," lab-credited) announcement on the Hodge Conjecture — only an unverified viral X post (2026-09-09) claiming OpenAI is "close." Note: OpenAI's Navier-Stokes claim (also 2026-09-09) does NOT qualify for this event per the rules (Navier-Stokes explicitly excluded). No qualifying announcement has occurred on any of the five eligible problems (BSD, Hodge, P vs NP, Riemann, Yang-Mills) as of this writing. # Timeline of key events - 2026-06/07 (reported): Prior AI math wins — OpenAI model reportedly "cracked" an Erdős conjecture; Claude found a counterexample to the Jacobian conjecture; an AI model formalized Fermat's Last Theorem in 11 days. Pattern of rapid, contested claims established. [claude_news/theweek] - 2026-07-23 (confirmed): Fields Medal awarded to Yu Deng, John Pardon, Jacob Tsimerman, Hong Wang — unrelated to AI/Millennium Problems but signals active mathematical progress landscape. [WaPo/NPR] - 2026-08-11 to 08-13 (reported): An unreleased Anthropic/Claude model improved lower bounds on the fraction of non-trivial zeta zeros on the critical line (Riemann Hypothesis) — a genuine but partial result, not a solution. [TechCrunch/TechSpot] - 2026-09-08/09 (confirmed): OpenAI announced ~10,000 agents worked 88 hours to produce a proposed solution/counterexample to Navier-Stokes; not yet verified by Clay Institute; ignited credit dispute with mathematician Tristan Buckmaster and Anthropic researcher Levent Alpöge. Navier-Stokes does NOT qualify for this market. [CNBC/CNN/Wikipedia/Science.org] - 2026-09-09 (rumored only): Viral X post claims Hodge Conjecture "very close to being verified by OpenAI," and that OpenAI or Anthropic close to BSD. No paper, no code, no institutional confirmation — explicitly labeled rumor by researchers. [vocal.media] # Event Will the Hodge Conjecture be the specific Millennium Prize Problem (among BSD, Hodge, P vs NP, Riemann, Yang-Mills) first solved and announced by an eligible AI lab by Dec 31, 2027? # Outcomes to forecast Yes / No (this ticker is the "Hodge" outcome within a larger multi-outcome event also containing BSD, P vs NP, Riemann, Yang-Mills, and "no solution by Dec 31, 2027") # Kalshi market anchor Kalshi-direct price not returned in this research pull; kalshi_related search found zero related markets. Polymarket (same event structure, likely cross-listed) shows the Hodge outcome at **41.5%** currently, up sharply from ~26.5% low, +15pts over both 7-day and 30-day windows (only 4 data points — thin/early history), volume ~$32k. This Polymarket price should be treated as the closest available consensus anchor in absence of direct Kalshi data. # Sub-question answers 1. **Polymarket price for Hodge vs siblings** — Hodge currently 41.5% per polymarket_direct (most recent), but claude_news (likely slightly earlier snapshot) cites BSD ~46%, Hodge ~30.5%, no-resolution ~13.5%. These are inconsistent/possibly different snapshots or different markets; Hodge has clearly risen recently (both agree on an uptrend), likely spurred by the Sept 9 rumor. 2. **Labs targeting Millennium Problems** — OpenAI publicly claimed Navier-Stokes solved (doesn't count here); Anthropic's unreleased model made Riemann progress (partial, not solved); no lab has officially named Hodge as a target — only unconfirmed rumor. No official Harmonic/Axiom Math/Math Inc statements surfaced in research. 3. **Current AI math capability** — Demonstrated wins: Erdős conjecture claim, Jacobian conjecture counterexample, FLT formalization in 11 days, Riemann zero-density bound improvement, and a large-scale agentic Navier-Stokes attempt. Capability is clearly advancing into research-frontier territory, but rigor/verification lags claims (Navier-Stokes still unverified days later). 4. **Hodge tractability vs other problems** — Mixed signal: one analysis argues algebraic geometry (Hodge's domain) may suit LLM strengths better than number theory (Riemann); but claude_news also states "BSD and Hodge have seen the least visible reinvigoration, with no major 2026 breakthroughs" — Riemann has the most concrete AI progress. No confirmed partial progress specifically on Hodge. 5. **Premature/contested claims precedent** — Yes, strong pattern: Navier-Stokes claim is disputed (credit dispute with Buckmaster/Alpöge, unverified by Clay Institute); prior Erdős/Jacobian claims fit same pattern of AI labs presenting results before independent verification. High propensity for premature announcements exists. 6. **Kalshi/other markets agreement** — kalshi_related returned zero matches; no independent Kalshi cross-check available. Polymarket is the only comparable market found, showing internal inconsistency between data pulls (30.5-46% for BSD-type ranking vs. 41.5% latest Hodge figure). # Key facts (high-confidence, factual) 1. [Wikipedia] Only Poincaré Conjecture has been officially solved (Perelman, 2010) among the seven Millennium Problems; Navier-Stokes claim is unverified/disputed as of Sept 2026. 2. [CNBC/CNN] OpenAI's Sept 9, 2026 Navier-Stokes announcement does not qualify for this market per explicit exclusion in rules. 3. [vocal.media] Hodge Conjecture "solution" is rumor-only — no paper, code, or institutional confirmation. 4. [TechCrunch/TechSpot] Anthropic's unreleased model improved a Riemann-related zero bound — a partial result, not a full solution, and on Riemann not Hodge. 5. [claude_news] Prior AI claims (Erdős, Jacobian counterexample, FLT formalization) show pattern of fast, contested announcements. # Cross-market signals - Kalshi related: none found. - Polymarket: Hodge outcome at 41.5% (latest), up from 26.5% low; other reporting suggests BSD may be seen as higher-probability alternative (~46% in an earlier/different snapshot), Hodge ~30.5%, no-solution ~13.5%. Directional trend for Hodge is upward, likely driven by unverified rumor. - Sportsbook implied: N/A (not applicable to this event type). # Analyst opinions and speculation - Terence Tao (CNN, Smithsonian): cautious/skeptical about AI math "solutions," warns of "meaningless production quota game," but also calls these problems "lighthouses" for research focus. - Market commentary (Polymarket analysis) suggests algebraic geometry (Hodge's domain) may suit current LLM strengths better than number theory-heavy Riemann — a structural argument favoring Hodge over Riemann, though not over BSD. - Rumor-driven speculation (viral X post) is unconfirmed and explicitly disputed by absence of any institutional or lab statement. # Directional lean per outcome - **Yes (Hodge)**: Supporting — recent price surge (+15pts), plausible structural fit for LLM strengths in algebraic geometry, pattern of labs rushing claims. Opposing — no actual paper/evidence, rumor explicitly unconfirmed, other sources suggest BSD or Riemann more likely given AI's demonstrated strengths in number theory/elliptic curves and concrete zero-bound progress. - **No (Hodge is not next solved)**: Supporting — reconciled evidence suggests Riemann has most tangible 2026 progress, BSD is analyst-favored in at least one dataset, Hodge/BSD described as having "least visible reinvigoration"; base rate of any single specific problem being solved by a specific date is low given multi-year unsolved status. # Gaps / unknowns - No direct Kalshi YES price retrieved for this ticker — reliance on Polymarket as proxy, which itself has internally inconsistent snapshots. - No official lab statements on Hodge specifically — only social media rumor with zero corroboration. - Unclear how "eligible AI lab" announcement bar (rigorous, lab-credited, "solved" language) will be applied given messy precedent (Navier-Stokes dispute). - No data on P vs NP or Yang-Mills mentioned at all in research — full multi-outcome distribution unclear. # Calibration anchors - Polymarket YES price for Hodge (proxy anchor): ~41.5% (volatile, thin data, 4 points, recent sharp rise). - Alternative/earlier snapshot cited: Hodge ~30.5%, BSD ~46%, no-solution ~13.5% — suggests real uncertainty/inconsistency in current pricing; treat 41.5% as upper/most recent estimate, but discount for rumor-driven volatility and small sample. - Precedent: Only 1 of 7 Millennium Problems solved in 26 years (Poincaré, 2010) — extremely low base rate for any single problem being solved in a ~1.3 year window, even with AI acceleration.

4. Forecasters Fable 5.1 + GPT-5.6 Sol · in parallel

claude-fable-5-1

30% conf

Yes	34%
No	66%

The only consensus anchor is Polymarket's Hodge price at ~41.5%, but it is thin (4 data points), internally inconsistent with an earlier snapshot (~30.5% Hodge vs ~46% BSD), and its +15pt surge is driven by an unverified viral X post with no paper, code, or lab statement. Competing outcomes have more tangible support: Anthropic's model made concrete partial progress on Riemann, BSD is favored in at least one market snapshot, and the 'no solution by Dec 2027' bucket is arguably underpriced given only one Millennium Problem solved in 26 years. Offsetting this, labs have demonstrated a strong propensity to announce contested 'solutions' quickly (Navier-Stokes), and the resolution bar is a lab announcement rather than Clay verification, so some specific problem plausibly gets 'claimed' in the window. I shade below the latest market price toward the average of the two snapshots, landing at 0.34 for Hodge being the first claimed among the five eligible problems.

gpt-5.6-sol

57% conf

Yes	30%
No	70%

The 41.5% Polymarket proxy is the best available consensus anchor, but its sharp rise appears driven largely by an uncorroborated September 9 rumor, warranting a substantial discount. No paper, code, official targeting statement, or confirmed Hodge-specific progress has surfaced, while Riemann has tangible partial AI progress and an earlier market snapshot favored BSD. Recent research-level AI results and labs' willingness to announce contested solutions keep Hodge meaningfully plausible, especially given the possible fit between algebraic geometry and current models. Nonetheless, the low historical solution rate, short deadline, competing eligible problems, and possibility of no qualifying solution make No more likely.

5. Devil's Advocate Sonnet 5 · red-team critique

Both forecasters converge tightly (0.30–0.34) despite explicitly flagging that the Polymarket "anchor" is thin (4 points), internally inconsistent (30.5% vs 41.5% Hodge snapshots), and rumor-driven — yet neither meaningfully widens their uncertainty or considers that the true anchor could be much lower (e.g., near the 13.5% "no-resolution" bucket implying Hodge specifically could be far under 30%). Given the brief's own gaps section ("No official lab statements on Hodge specifically"), a probability in the low-20s or even teens seems equally defensible as 30-34%.

Neither forecast grapples with the base-rate anchor explicitly given in the brief: "Only 1 of 7 Millennium Problems solved in 26 years" — this argues for a much lower probability on *any* problem being solved by Dec 2027, which should pull down Yes probabilities for all five sub-outcomes, not just redistribute mass toward BSD/Riemann while leaving Hodge at ~30%+. Both forecasts effectively ignore how this base rate should compress the entire distribution rather than just reshuffle relative rankings.

Both forecasters lean on "labs have shown propensity to announce contested solutions" as support for Yes, but neither addresses the specific asymmetry that Hodge is explicitly the *least* evidenced of the five problems ("BSD and Hodge have seen the least visible reinvigoration, with no major 2026 breakthroughs") — this should push Hodge's probability of being the *specific* first-claimed problem below Riemann/BSD by more than the ~5-10pt gap implied in their numbers, especially since the algebraic-geometry-fit argument is speculative commentary, not evidence of actual work in progress.

Neither forecast addresses P vs NP or Yang-Mills at all, which the brief flags as a complete gap ("No data on P vs NP or Yang-Mills mentioned at all in research — full multi-outcome distribution unclear") — if these carry non-trivial probability mass, Yes(Hodge) should likely be lower than both forecasts suggest, since the five-way (plus no-solution) split isn't just Hodge/BSD/Riemann/no-solution as implicitly modeled.

Both treat the viral rumor and the Navier-Stokes precedent as roughly interchangeable evidence for "labs rushing announcements," but the Navier-Stokes case is a confirmed announcement (contested but real), whereas the Hodge rumor has zero corroboration even at rumor-source level (no paper/code/institutional statement) — this qualitative difference in evidence tier is somewhat glossed over, and arguably justifies a probability meaningfully below the Polymarket 41.5% headline, closer to Forecast 2's 0.30 than Forecast 1's 0.34.

Neither forecaster explicitly discusses resolution-criteria ambiguity noted in the brief ("Unclear how 'eligible AI lab' announcement bar... will be applied given messy precedent") — if the bar is genuinely ambiguous or contested (as with Navier-Stokes), this could push toward "No" resolving more often across all problems (nothing clearly qualifies), which neither forecast incorporates as a distinct driver toward lower Yes probabilities.

6. Aggregator Fable 5.1 · final calibrated forecast

32% (43% confidence)

Yes: 32% No: 68%

The only market anchor is a thin, rumor-driven Polymarket price (41.5% latest vs ~30.5% in an earlier snapshot), so averaging the snapshots and discounting for the uncorroborated Sept 9 X post is appropriate. Hodge has zero confirmed AI progress while Riemann has tangible partial results and BSD was market-favored in one snapshot; P vs NP and Yang-Mills also carry unmodeled mass, and resolution-criteria ambiguity plus the historical base rate argue for a fatter 'no qualifying solution' bucket. Offsetting this, labs (especially OpenAI, with its 10,000-agent Navier-Stokes effort) have shown they will announce contested 'solutions' quickly, the bar is a lab announcement rather than Clay verification, and the specific Hodge rumor plausibly reflects real OpenAI interest that the market is pricing. Both forecasters drew the same inference and the critique mostly reinforces shading below the headline price rather than exposing a shared flaw, so I settle at 0.32, between the two forecasts and below the latest market print.

7. Entry Decision Fable 5.1 trader

Side

no

Position Size

$600

Contracts

1017

Trader Confidence

44%

Trade rationale

Forecast 32% vs market 41.5% gives a -9.5pp edge pointing NO. This is a direction-agreement setup (both sides below 50%), so the bar is higher, but the rationale identifies something concrete and recent: the price jumped from ~30.5% to 41.5% on an uncorroborated Sept 9 X post in a thin $32k market, which looks like a rumor-driven overshoot rather than processed information. Critically, the Devil's Advocate does not undermine the NO thesis — every one of its four points argues the probability should be even lower (1-of-7-in-26-years base rate, Hodge being the least-evidenced problem, unmodeled mass on P vs NP and Yang-Mills), so the critique strengthens the trade direction. Forecaster confidence is 0.43 with a tight 4pp ensemble spread; the main risk is that labs announce contested 'solutions' quickly and Hodge's algebraic-geometry fit attracts a real OpenAI effort, plus 475 days of horizon.

Allocation logic

Sized at $600, the low end, because this is a sub-15pp direction-agreement edge (historically ~36% hit rate), the market is thin, the horizon is long, and the book already has several AI-lab/math-AI positions (Meta #3 lab, Google math AI, Kimi) that share theme exposure.

Entry price: $0.59

Current: $0.69

Status: OPEN

P&L: $91.85

Pipeline Timing

Total pipeline time: 298.2s

Per-tool research timings shown in the Research section above.