How many more Millennium Prize Problems will AI solve in 2026?
The most likely outcome is that no qualifying Millennium Prize Problem is formally announced as solved by an eligible AI lab in 2026. The headline Navier-Stokes claim does not count for this market, and the remaining window for a clean qualifying announcement is still short.
Analysis
The central point for this market is that the widely discussed OpenAI Navier-Stokes claim does not qualify, because this market explicitly excludes announcements concerning Navier-Stokes. That means the biggest recent AI-math headline does not move the resolution needle directly, even if it has created a strong impression that a breakthrough has already happened. For the market to resolve to No, it only needs there to be no qualifying announcement on the other five problems by the end of 2026, and so far there is no clear public evidence of that happening.
The recent reporting does, however, support the idea that AI labs are pushing closer to major results, which lowers certainty around a zero-solve year. A later report that OpenAI discussed substantial progress on another Millennium Prize Problem suggests the field may be closer to producing additional claims, even if those claims are not yet public, not yet identified, or not yet framed as solved. That creates some real upside risk for Yes, because this market keys off announcements from eligible labs rather than formal certification, so a public declaration from a qualifying lab could arrive quickly and settle the market against No.
Still, the base rate strongly favors no qualifying announcement in the remaining months of 2026. These problems are extraordinarily hard, and even in a year with intense AI research activity, a credible announcement for one of the five qualifying problems would be a rare event. The most plausible path to Yes is an abrupt, high-confidence announcement from a major lab on one of the unsolved problems, while the most plausible path to No is simply that the current excitement remains limited to partial progress, disputed claims, or work on the excluded Navier-Stokes problem.
Market pricing near 50/50 seems to reflect uncertainty about how to interpret the recent news flow, but I think that is too high for a true qualifying announcement by year-end. Because the known Navier-Stokes claim is non-qualifying and no other eligible lab has publicly announced a solved qualifying problem, the default expectation should still be that 2026 ends with zero counted solves. The main reason to stay below an extreme conviction level is that a single surprise announcement from one eligible lab would immediately flip the outcome.
Arguments
For
- Arguments for Yes: The recent AI-math headlines suggest a real chance that at least one qualifying Millennium Prize Problem may be publicly claimed as solved before year-end.
- Arguments for Yes: The market could react quickly to a surprise announcement from a major lab, especially if the claim is framed unambiguously as a full solution.
Against
- Arguments against Yes: The most prominent reported breakthrough is on Navier-Stokes, and that problem is explicitly excluded from counting.
- Arguments against Yes: There is no publicly confirmed qualifying announcement for any of the other five problems, and a zero-solve year remains the default outcome absent a new claim.
Key drivers
- The only clearly public AI math breakthrough mentioned so far is Navier-Stokes, which is explicitly excluded from qualification.
- There is still enough time left in 2026 for a surprise announcement, but the window is shrinking and the bar for a qualifying claim is very high.
Risk factors
- An eligible lab could announce a solution to one of the five qualifying problems on short notice.
- Market participants may be overconfident that official recognition is required, when the rules also allow qualifying lab announcements without outside certification.
Scenarios
Best case
An eligible AI lab publicly announces a convincing solution to one of the five qualifying problems before December 31, causing the market to resolve to No.
Most likely
The Navier-Stokes story continues to generate headlines but does not count for this market, and no qualifying announcement appears before year-end, so the market resolves to Yes.
Worst case
No eligible lab announces a solution to any of the five qualifying problems, and the market resolves to Yes on zero more problems in 2026.
More from this day
- PoliticsKalshi2y
Which agencies will Trump eliminate?
AI87%MKT19%Edge+68Hidden GemI assess a high likelihood that the market’s Yes outcome occurs, because the available evidence suggests USAID has already been functionally eliminated during Trump’s term and remaining obstacles look mostly legal or semantic rather than operational. The main uncertainty is whether the market resolves “eliminated” as formal statutory repeal versus de facto shutdown and absorption into State.
- PoliticsKalshi1y
Which state will vote first in the 2028 Democratic presidential primary?
AI33%MKT89%Edge-56HypedThe evidence favors New Hampshire not being the first Democratic contest in 2028. Recent calendar reporting points to South Carolina first, with New Hampshire early but not earliest.
- pop culturePolymarketEnded
#2 Spotify song in the US this week? (September 18)
AI38%MKT87%Edge-49HypedBbY WOW has strong momentum and could plausibly land at number two, but the U.S. streaming evidence still suggests it is more likely to sit just behind a stronger domestic leader rather than clearly control the runner-up spot. I would price a moderate chance rather than a favorite outcome.