# Will NASDAQ fall at least 65% from its historical peak at any time before the end of 2030?

Canonical URL: https://preseen.com/reports/efe5c48d-1b2a-4dee-b1ff-0c5d55dcce71/will-nasdaq-fall-at-least-65-from-its-historical-peak-at-any-time-before-the-end
Markdown URL: https://preseen.com/reports/efe5c48d-1b2a-4dee-b1ff-0c5d55dcce71/markdown

## Forecast

P(Yes): 10.1%; P(No): 89.9%.

Generated: September 10, 2026 at 1:41 AM UTC
Forecast model: gpt-5.6-sol
Research model: gpt-5.6-sol

## Analysis

## TL;DR
My forecast is **10% YES**. The threshold is extreme: the Nasdaq Composite has crossed it in only one fresh-record drawdown since the [Nasdaq-sourced daily-close history](https://fred.stlouisfed.org/series/NASDAQCOM) began in 1971, during the dot-com collapse. High concentration and record-scale AI spending raise the tail, but strong profits and calm credit markets make NO the clear base case. ([]())

## Context
The [Nasdaq Composite](https://indexes.nasdaq.com/Index/Overview/COMP) closed at 26,253.34 on September 9, 2026. Its running all-time closing high was 27,093.90 on June 2, 2026, making the current drawdown 3.10%. The present trigger is 9,482.865, or another 63.88% below the September 9 close. ([indexes.nasdaq.com](https://indexes.nasdaq.com/Index/Overview/COMP))

That trigger is not fixed. A new record would raise the absolute qualifying level to 35% of the new peak. The remaining horizon is about 4.31 years, so the question covers both a direct collapse from today’s level and a further boom followed by a crash.

## Evidence
The historical backbone is the full [FRED Nasdaq Composite series](https://fred.stlouisfed.org/series/NASDAQCOM): 14,015 numeric official closes from February 5, 1971 through September 8, 2026, followed by the September 9 close on [Nasdaq’s live page](https://indexes.nasdaq.com/Index/Overview/COMP). The table lists every distinct running-all-time-peak episode that fell at least 20%. The 2007–09 decline is not a separate row because the index was still underwater from its 2000 record. ([]())

| Fresh running-peak episode | Peak to trough | Maximum closing drawdown |
|---|---|---:|
| [1973–74](https://fred.stlouisfed.org/series/NASDAQCOM) | January 11, 1973 → October 3, 1974 | 59.90% |
| [1978](https://fred.stlouisfed.org/series/NASDAQCOM) | September 13, 1978 → November 14, 1978 | 20.37% |
| [1980](https://fred.stlouisfed.org/series/NASDAQCOM) | February 8, 1980 → March 27, 1980 | 24.91% |
| [1981–82](https://fred.stlouisfed.org/series/NASDAQCOM) | May 29, 1981 → August 13, 1982 | 28.79% |
| [1983–84](https://fred.stlouisfed.org/series/NASDAQCOM) | June 24, 1983 → July 25, 1984 | 31.50% |
| [1987](https://fred.stlouisfed.org/series/NASDAQCOM) | August 27, 1987 → October 28, 1987 | 35.96% |
| [1989–90](https://fred.stlouisfed.org/series/NASDAQCOM) | October 9, 1989 → October 16, 1990 | 33.00% |
| [1998](https://fred.stlouisfed.org/series/NASDAQCOM) | July 20, 1998 → October 8, 1998 | 29.55% |
| [2000–02](https://fred.stlouisfed.org/series/NASDAQCOM) | March 10, 2000 → October 9, 2002 | 77.93% |
| [2018](https://fred.stlouisfed.org/series/NASDAQCOM) | August 29, 2018 → December 24, 2018 | 23.64% |
| [2020](https://fred.stlouisfed.org/series/NASDAQCOM) | February 19, 2020 → March 23, 2020 | 30.12% |
| [2021–22](https://fred.stlouisfed.org/series/NASDAQCOM) | November 19, 2021 → December 28, 2022 | 36.40% |
| [2024–25](https://fred.stlouisfed.org/series/NASDAQCOM) | December 16, 2024 → April 8, 2025 | 24.32% |

Only the dot-com episode crossed 65%. It fell from 5,048.62 to 1,114.11 and first closed below the 65% threshold on April 3, 2001. The closest other episode was 1973–74 at 59.90%. Even the global financial crisis produced only a 55.63% decline from its October 2007 local high. These near-misses matter because they show how much more severe the resolution threshold is than a normal recession or bear market. ([]())

One qualifying fresh-peak episode in roughly 55.6 years gives a simple 4.31-year Poisson estimate of `1 − exp(−4.31 / 55.6) = 7.5%`. A correctly aligned non-overlapping-window calculation gives one event in 12 windows, or 8.3%. Bayesian versions range from about 6% to 11%, depending on how strongly the prior favors the lower crash rates seen in broader equity indexes.

Return-path models form a lower band. Block bootstraps and Student-t GARCH models that update the running maximum produce roughly 4%–8% under reduced or zero forward drift. Full historical drift drives estimates below 3%, but that assumes the Nasdaq keeps compounding near its long-run average while also treating structural bubbles as ordinary return sequences. I give those low estimates little weight. The disagreement between models is itself evidence that the far tail is driven by regime changes, not by routine daily volatility.

Current conditions justify moving above the raw historical rate. The [June 30 Nasdaq fact sheet](https://indexes.nasdaqomx.com/docs/FS_COMP.pdf) put the ten largest securities at 53.29% of COMP and technology at 63.51%. Nvidia and Apple alone were close to one-fifth of the index. The Federal Reserve’s [July meeting minutes](https://www.federalreserve.gov/monetarypolicy/fomcminutes20260729.htm) said the equity premium had only been lower in recent history during the dot-com bubble and described hedge-fund leverage as near record highs. ([indexes.nasdaqomx.com](https://indexes.nasdaqomx.com/docs/FS_COMP.pdf))

AI capital spending supplies a plausible crash mechanism. [FactSet](https://insight.factset.com/hyperscalers-tap-external-financing-as-ai-capex-outruns-cash-flow) estimated fiscal-2026 capex above $690 billion for five hyperscalers, more than 80% higher year over year; incremental debt had risen from 9% of capex in fiscal 2024 to 32% by mid-2026. Direct disclosures show first-half cash capex of [$96.3 billion at Amazon](https://www.sec.gov/Archives/edgar/data/1018724/000101872426000026/amzn-20260630.htm) and [$80.6 billion at Alphabet](https://www.sec.gov/Archives/edgar/data/1652044/000165204426000071/goog-20260630.htm), while [Meta projected $130–145 billion](https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/meta-20260630.htm) for 2026. [Microsoft](https://www.microsoft.com/en-us/investor/events/fy-2026/earnings-fy-2026-q4) spent $41 billion in its latest quarter, with roughly two-thirds going to shorter-lived CPUs and GPUs. 

This is not yet a solvency crisis. FactSet reported debt-to-EBITDA around one times or less for four of the five hyperscalers. [Nasdaq’s second-quarter earnings review](https://www.nasdaq.com/economic-institute/the-print/second-quarter-earnings-show-breadth) found that more than half of Nasdaq-100 companies had earnings growth above 20%, and more than four-fifths had positive growth. Some of that result reflects unusual gains, and Nasdaq has an institutional bias toward a favorable reading, but the core point holds: today’s largest companies have real revenue, cash flow and profits. 

The macro data also argues against an imminent cascade. August payrolls rose 162,000 and unemployment was 4.1% in the [September 4 BLS release](https://www.bls.gov/news.release/archives/empsit_09042026.htm). The [real-time Sahm indicator](https://fred.stlouisfed.org/series/SAHMREALTIME) was −0.07 percentage point, the [NFCI](https://fred.stlouisfed.org/series/NFCI) was −0.558, and the [high-yield spread](https://fred.stlouisfed.org/series/BAMLH0A0HYM2) was 2.67 percentage points. Those readings show loose financial conditions, not recessionary credit stress. The Fed also described banks as sound and resilient in its [May Financial Stability Report](https://www.federalreserve.gov/publications/2026-may-financial-stability-report-overview.htm). ([bls.gov](https://www.bls.gov/news.release/archives/empsit_09042026.htm))

The offset is that policymakers have less freedom than those calm indicators suggest. July headline and core PCE inflation were 3.7% and 3.3% year over year in the [August 26 BEA release](https://www.bea.gov/news/2026/personal-income-and-outlays-july-2026). The [10-year Treasury yield](https://fred.stlouisfed.org/series/DGS10) was 4.80% on September 8. Three FOMC members preferred a rate increase in July. A growth shock with stubborn inflation would therefore be more dangerous than a standard disinflationary recession, because rapid monetary easing could be delayed. ([bea.gov](https://www.bea.gov/news/2026/personal-income-and-outlays-july-2026))

I start from a historical anchor near 7.5%. Concentration, thin valuation compensation, AI investment intensity and limited inflation-adjusted policy room justify about a 1.4-times increase in the event odds. Strong earnings, low mega-cap net leverage, healthy banks and tight credit spreads keep the adjustment from being larger. This maps to a central estimate close to 10%, and a separate scenario analysis based on ordinary recession, severe AI retrenchment and systemic-shock pathways lands in the same range.

## What's non-obvious
The global financial crisis can easily be miscounted as a second historical 65% event. In March 2009 the Nasdaq was more than 65% below its still-unrecovered March 2000 record, but the decline from the October 2007 local high was only 55.63%. Treating overlapping windows or arbitrary calendar blocks as independent can therefore turn one long 2000–15 underwater period into two apparent crashes and inflate the forecast. [The daily series](https://fred.stlouisfed.org/series/NASDAQCOM) supports one fresh-record event, not two. ([]())

AI is also not a simple bearish input. The spending creates current revenue and earnings for chip, cloud, power and construction firms. The risk comes later if capacity is overbuilt, pricing falls, or short-lived hardware must be replaced before its economic return is clear. Accounting depreciation and financing costs lag the cash outlay, so strong present earnings and rising future vulnerability can coexist. The data shows a real boom with real profits, not a repeat of 2000; my inference is that this still raises the chance of a rare joint earnings-and-valuation collapse.

## Uncertainties
The direct sample contains one qualifying event. Priors, drift assumptions and the treatment of the 2000–15 underwater period can move a statistical estimate by several percentage points. The index itself has also changed: today it is more concentrated in profitable mega-caps, but more exposed to one technology and capital-spending cycle.

The official [FRED series](https://fred.stlouisfed.org/series/NASDAQCOM) was one session behind at the cutoff, so the September 9 close came from [Nasdaq’s live page](https://indexes.nasdaq.com/Index/Overview/COMP). The latest publicly accessible official component-weight snapshot was June 30, not September 9. There is no standardized measure of AI-only capex, and companies differ in their treatment of finance leases, operating leases and cash purchases. ([]())

My reasonable probability range is 5%–17%. The lower end applies if AI revenue keeps validating the spending and credit stays healthy. The upper end applies if capex increasingly relies on debt and leases, earnings estimates reverse, unemployment rises, and high inflation or long-term rates block a fast policy response.

## Sources

- Domain Expert Search (mcp)
  > Found 14 domain experts for 'Nasdaq Composite crash risk, equity market drawdowns, valuation bubbles, AI capital expenditure, US financial stability through 2030':
- FRED (mcp)
  > Series: NFCI
- [Chicago Fed National Financial Conditions Index (NFCI) | FRED | St. Louis Fed](https://fred.stlouisfed.org/series/NFCI) (openai)
- [ICE BofA US High Yield Index Option-Adjusted Spread (BAMLH0A0HYM2) | FRED | St. Louis Fed](https://fred.stlouisfed.org/series/BAMLH0A0HYM2) (openai)
- [Market Yield on U.S. Treasury Securities at 10-Year Constant Maturity, Quoted on an Investment Basis (DGS10) | FRED | St. Louis Fed](https://fred.stlouisfed.org/series/DGS10) (openai)
- [Real-time Sahm Rule Recession Indicator (SAHMREALTIME) | FRED | St. Louis Fed](https://fred.stlouisfed.org/series/SAHMREALTIME) (openai)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/UNRATE) (tool)
- Cboe vix (mcp)
  > VIX as of 2026-09-09:
- Clevelandfed (mcp)
  > Cleveland Fed Yield Curve Recession Probability (Monthly)
- Yfinance (mcp)
  > Mode: specific dates lookup | Dates requested: 2
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/USREC) (tool)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/DFF) (tool)
- Domain Expert Research Task (mcp)
  > Job domain_expert_research_task_ce04a10288 done after 548144ms.
- [NASDAQ Composite (NASDAQCOM) | FRED | St. Louis Fed](https://fred.stlouisfed.org/series/NASDAQCOM) (openai)
- [nber.org](https://www.nber.org/news/business-cycle-dating-committee-announcement-july-19-2021) (tool)
- [nber.org](https://www.nber.org/research/data/us-business-cycle-expansions-and-contractions) (tool)
- [bea.gov](https://www.bea.gov/news/2026/gdp-second-estimate-and-corporate-profits-2nd-quarter-2026) (tool)
- [Personal Income and Outlays, July 2026 | U.S. Bureau of Economic Analysis (BEA)](https://www.bea.gov/news/2026/personal-income-and-outlays-july-2026) (openai)
- [federalreserve.gov](https://www.federalreserve.gov/monetarypolicy/files/BeigeBook_20260902.pdf) (tool)
- [Employment Situation News Release - 2026 M08 Results](https://www.bls.gov/news.release/archives/empsit_09042026.htm) (openai)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/ICSA) (tool)
- [bls.gov](https://www.bls.gov/news.release/jolts.htm) (tool)
- [bls.gov](https://www.bls.gov/news.release/archives/prod2_09032026.htm) (tool)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/BAMLC0A0CM) (tool)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/ANFCI) (tool)
- [St. Louis Fed Financial Stress Index (STLFSI4) | FRED | St. Louis Fed](https://fred.stlouisfed.org/series/STLFSI4) (openai)
- [federalreserve.gov](https://www.federalreserve.gov/data/sloos/sloos-202607.htm) (tool)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/BUSLOANS) (tool)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/CCLACBW027SBOG) (tool)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/DGS2) (tool)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/T10Y2Y) (tool)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/DFII10) (tool)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/T10YIE) (tool)
- [federalreserve.gov](https://www.federalreserve.gov/newsevents/pressreleases/monetary20260729a.htm) (tool)
- [federalreserve.gov](https://www.federalreserve.gov/monetarypolicy/2026-07-mpr-summary.htm) (tool)
- [federalreserve.gov](https://www.federalreserve.gov/publications/2026-june-supervision-and-regulation-report-report-banking-system-conditions.htm) (tool)
- [federalreserve.gov](https://www.federalreserve.gov/publications/2026-may-financial-stability-report-funding-risks.htm) (tool)
- [The Fed - Overview](https://www.federalreserve.gov/publications/2026-may-financial-stability-report-overview.htm) (openai)
- [federalreserve.gov](https://www.federalreserve.gov/publications/2026-may-financial-stability-report-leverage.htm) (tool)
- [federalreserve.gov](https://www.federalreserve.gov/publications/2026-may-financial-stability-report-accessibility-tables.htm) (tool)
- [federalreserve.gov](https://www.federalreserve.gov/econres/notes/feds-notes/private-credit-and-leveraged-loan-markets-similarities-differences-and-substitution-20260811.html) (tool)
- [The Fed - 5. Near-Term Risks to the Financial System](https://www.federalreserve.gov/publications/2026-may-financial-stability-report-near-term-risks.htm) (openai)
- [federalreserve.gov](https://www.federalreserve.gov/publications/2026-may-financial-stability-report-asset-valuations.htm) (tool)
- [bls.gov](https://www.bls.gov/schedule/2026/home.htm) (tool)
- [www.finra.org](https://www.finra.org/rules-guidance/key-topics/margin-accounts/margin-statistics) (openai)
- Claude Code (e2b)
  > Job coding_whiz_job_06c7d4d977 done after 502775ms.
- [Overview for COMP](https://indexes.nasdaq.com/Index/Overview/COMP) (openai)
- [indexes.nasdaqomx.com](https://indexes.nasdaqomx.com/docs/FS_COMP.pdf) (openai)
- Oecd cci (mcp)
  > OECD Composite Leading Indicator (LI)
- [fred.stlouisfed.org](https://fred.stlouisfed.org/series/JTSJOL) (tool)

## Question Details

This question asks whether the Nasdaq Composite Index (ticker: COMP) will experience a drawdown of at least 65% from its highest official closing value ever recorded, with the drawdown occurring at any point on or before 2030-12-31. As of July 2026, the Nasdaq Composite is trading near all-time highs after setting new records during 2025-2026. The reference peak is the highest official closing value the Nasdaq Composite has achieved at any time in its history up to the moment the qualifying drawdown is evaluated. If a new all-time closing high is established before 2030-12-31, that new closing high becomes the reference peak for subsequent evaluation.

### Resolution Criteria

Resolve YES if, at the close of any trading day on or before 2030-12-31, the official closing value of the Nasdaq Composite Index is at least 65.0% below the highest official closing value the index has reached up to that date. Equivalently, let P be the highest official closing value recorded up to a given trading day, and let C be the official closing value on that trading day. Resolve YES if C <= 0.35 × P. Otherwise, resolve NO. The primary resolution source will be Nasdaq's official historical index data. If unavailable, FRED (Federal Reserve Bank of St. Louis, sourced from Nasdaq) or another widely recognized historical market data provider reporting official Nasdaq Composite closing values may be used.

### Fine Print

The question refers to the Nasdaq Composite Index, not the Nasdaq-100 Index or any ETF. Only official closing values are used. Intraday highs or lows that are not reflected in the official closing value do not count. Index revisions or corrections published by the official source before resolution should be used. If trading is halted or the market closes early, the officially published closing value for that trading session will be used.