The Dow closes above 1,000,000

Where things stand

• Current level: The DJIA closed at 52,002.41 on 2026-10-16 (FRED DJIA CSV, https://fred.stlouisfed.org/graph/fredgraph.csv?id=DJIA; matches Yahoo ^DJI). The close on the opening date (2026-09-16) was 51,461.90, and the 52-week range is 45,057 to 54,744.
• What YES needs: a gain of 19.23 times in about 5.2 years, which works out to a 76% compound annual rate to the end of 2031. Earlier quarter-ends need even faster growth: about 1,057%/yr by the end of 2027 and 151%/yr by the end of 2029. I computed these from the FRED/Yahoo close.
• Index changes don't help. The question converts any rebasing back to the 2026 scale, and constituent or divisor changes keep the series continuous. Alphabet replacing Verizon in June 2026 changed nothing about the level.

Base rate (computed)

• Long US history: I used Shiller's monthly S&P Composite, 1871–2023 (http://www.econ.yale.edu/~shiller/data/ie_data.xls). The largest nominal rise to a later peak within any window of 63 months or less is 3.80 times (June 1932 to February 1937). Within 12 months the maximum is 2.24 times. The 99.9th percentile of 5-year ratios is 3.3 times.
• Recent DJIA: On daily DJIA data since 1992 (Yahoo), the largest rise to a later peak within 5.2 years is 3.19 times.
• Elsewhere: Rises of about 20 times or more in under 6 years happened only in small, fast-growing markets (Taiwan 1985–90, Hang Seng 1967–73) or high-inflation economies (Argentina, Turkey, Zimbabwe). They never happened in a dominant market of mega-cap companies. The most extreme cases there were about 5 times (Nasdaq 1995–2000, Nikkei in the 1980s).
• Option-style model: A lognormal model with 17–20% volatility gives roughly 1e-10. So any real probability comes from regime breaks, not ordinary market moves.

Paths to YES

1. AI-driven explosive growth with a huge rerating of stocks (about 0.15–0.2%).
   • Current evidence points the other way. The 10-year inflation-protected (TIPS) real yield is about 2.8% (FRED DFII10). Economist Basil Halperin's rule of thumb is that real rates rise roughly one point for each point of extra growth, so this is not a singularity price.
   • When OpenAI released "Astra" and Jensen Huang declared AGI in September 2026, markets barely moved, and the AI trade is now "reeling" (Fortune, 2026-09-09; CNBC, 2026-10-15). Halperin's own forecast is "dot-com boom… twice as fast," which implies a few-fold rise, not 19-fold.
   • The Dow is price-weighted. The big AI-exposed names (Apple, Microsoft, Nvidia, Amazon, Alphabet) make up only about 22% of its weight together with Salesforce, IBM and Cisco (TechTimes, 2026-06-25). That means most old-economy components would also have to rise about 15–20 times.
   • Even in a fast takeoff, nominal stock prices rising 19 times by 2031 would need explosive growth to be well under way by about 2029–30.
2. US hyperinflation (about 0.05–0.1%).
   • Inflation is 3.7% year on year, driven by energy during the war with Iran, and the Fed is raising rates: it hiked in September and more hikes are priced in (CNBC, 2026-10-14).
   • Five-year breakeven inflation is 2.37% and the 5-year-5-year forward rate is 2.33% (FRED T5YIE and T5YIFR).
   • In advanced economies, inflation above 50% has occurred in only a handful of country-years since 1950. Manifold's 13% for US hyperinflation by 2030 reflects the known upward bias of long-dated, thinly traded markets, not credible information.
3. Data errors and quirks (about 0.01%). The administrator's value overrides a conflicting FRED value, and a bad print would be corrected.

Timing

The hazard is close to zero through 2027, because no plausible process produces 10 to 1,000% annual gains straight away. It then rises into 2030–31, when an extreme scenario would have had time to play out. The cumulative probability is about 0.003% by the end of 2027, 0.06% by the end of 2029 and about 0.27% by the end of 2031. I built this from quarterly hazards and their cumulative product.

Strongest case against my estimate

• For a higher number: people with short AI timelines might give 20–30% to superintelligence by 2029, with large stock-price effects, which would imply over 1%. I discount this because markets, real rates and the Dow's price-weighted, old-economy makeup all argue against it, and because real rates should rise under explosive growth, which caps valuations.
• For a lower number: with no historical precedent and option-implied odds near zero, the true figure could be about 0.05%. The difference is immaterial for scoring.

Other context: Mainstream analysts doubt even Dow 100,000 by 2030 (Northeastern, 2026-02-24). Buffett's "Dow 1 million" call was for around 2117 (CNBC, 2017).

Question. By the last day of each quarter, has the Dow Jones Industrial Average closed above 1,000,000 on any day?

Definitions.

• Dow Jones Industrial Average (DJIA): S&P Dow Jones Indices’ headline USD price-return index, measured in nominal index points, not its total-return version or an inflation-adjusted series.
• Close: the official daily closing index value for a trading session, including an officially shortened session. Intraday highs, futures prices, after-hours indications, and prices of funds tracking the index do not count.
• Above 1,000,000: strictly greater than 1,000,000 index points; a close of exactly 1,000,000 does not count. No attribution to AI or adjustment for inflation is required.
• Ordinary constituent and divisor changes count as part of the same series. If the index is explicitly rebased or its units redenominated, convert subsequent values to the scale in use on 2026-09-16 using the administrator’s published conversion factor. A renamed continuation counts; a distinct replacement index does not.

Resolves YES if a qualifying source publishes an official daily close strictly above 1,000,000 for a trading day ending after the opening date, by the quarter-end. The resolution window and observation-period conventions apply. An official correction that first establishes a qualifying close counts on its publication date, not retroactively. Once YES, always YES under the resolution-window convention.

Resolves NO if no qualifying closing value has been published by the quarter-end. Estimates, forecasts, hypothetical backtests, and revisions solely to observations ending on or before the opening date do not qualify.

Resolving source. S&P Dow Jones Indices’ Dow Jones Industrial Average index page, its official daily index data or announcements, or the Federal Reserve Bank of St. Louis’s FRED series DJIA, which republishes S&P Dow Jones Indices data with frequency “Daily, Close.” The administrator’s value takes precedence over a conflicting FRED value. Use the publication date per convention A3, not merely the observation date attached to a subsequently published value. These named statistical sources govern; public confirmation under A2 is not required.