Will NASDAQ fall at least 65% from its historical peak at any time before the end of 2030?
Latest forecast
Generated Nov 7, 2026, 1:52 AM
Write-up
Nasdaq Composite: risk of a 65% drawdown by 2030
Routine update
The update adds evidence on both AI financing vulnerabilities and valuation resilience, without a clear explanation for the lower forecast. The chance of a 65% Nasdaq Composite closing drawdown by end-2030 remains about 11%, down less than one percentage point; avoiding that extreme decline is still the clear base case.
Executive Summary
The probability that the question resolves YES is 10.6%. A qualifying collapse is a meaningful risk, but NO remains the clear central expectation. The principal danger is a prolonged AI investment bust that spreads from earnings into customer financing and credit markets—not simply a correction in AI-related shares.
At the latest completed session, November 6, 2026, the Nasdaq Composite closed at 24,579.42, against a running closing record of 27,599.89 on October 6 in Nasdaq-sourced FRED data. Without a subsequent higher peak, the qualifying close would be 9,659.9615 or lower, requiring another 60.70% decline from the latest close. This is substantially more severe than an ordinary bear market.
The forecast treats the question prospectively from the client’s July 21, 2026 submission through December 31, 2030. No qualifying close occurred through November 6. Only official Nasdaq Composite closing values count; each new record closing high raises the reference peak, and a single qualifying close would settle YES even if the index subsequently recovered.
Key drivers
Historical rarity is the strongest restraint on the forecast
A drawdown is the percentage decline from the highest closing value reached so far. Across the daily Nasdaq Composite record beginning in 1971, only one independent record-peak-to-recovery episode exceeded the question’s threshold: the dot-com collapse, whose deepest closing decline was 77.93%.
Other major downturns stopped short. The 1973–1974 decline reached 59.90%, while the 2007–2009 financial crisis produced a 55.63% decline from its local peak. The latter also generated qualifying closes against the still-unrecovered 2000 record, but it was not a second independent crash from a new all-time high. That distinction prevents repeated observations of the same long episode from overstating the historical evidence.
AI investment is becoming more dependent on financing
The main reason not to rely on historical rarity alone is the scale and funding structure of the AI buildout. For July–September 2026, Microsoft, Amazon, Alphabet and Meta collectively spent more on cash property and equipment than they generated in operating cash flow. The calculated quarterly shortfall was $25.04 billion.
That comparison excludes finance-lease principal and nets Amazon’s sales and incentives against purchases. It is neither an AI-only spending measure nor a standardized industry free-cash-flow total. Nevertheless, it shows increasing dependence on reserves, external financing or stronger future cash generation. Because the Composite is capitalization-weighted—larger companies have greater influence—stress among its leading technology firms can matter disproportionately.
Financial connections could turn disappointment into contagion
Amazon disclosed a $50 billion investment in OpenAI alongside major commercial cloud arrangements. Nvidia disclosed a capped $105 billion SB Energy/OpenAI-related guarantee, phased with construction and lease commencement, with the first phase expected in fiscal 2029. These are different exposures and should not be added together; the guarantee is contingent support, not immediately payable debt.
Their significance is the connection between financing customers and relying on those customers for future demand. A disappointment could therefore damage sales, investment plans and credit exposure together. Meta’s non-cancelable contracts and leases not yet commenced also weaken the assumption that all infrastructure spending could be stopped quickly.
Strong operating businesses and a resilient economy remain important counterweights
The leading companies generate substantial current earnings. A cash deficit after heavy investment is not equivalent to an unprofitable underlying business.
Valuation evidence also cuts against an indiscriminate bubble comparison. Nasdaq’s October 30 dashboard reported a Nasdaq-100 forward price-to-earnings ratio—price relative to expected earnings—of 20.30 as of October 28, below its ten-year average of 23.10. This is useful counter-evidence, but it concerns the Nasdaq-100, not the Composite, and depends on forecast rather than recession-normalized earnings.
The economic starting point was not a broad collapse: October unemployment was 4.1%, and credit markets remained open. Elevated inflation, however, limits the assumption that monetary policy could freely cushion a future shock.
Scenarios
The forecast uses three mutually exclusive scenarios. Their relative weights reflect judgment informed by the historical record and current financial conditions, rather than mechanically fitted probabilities.
| Scenario | Relative weight and rationale | How it leads to an outcome |
|---|---|---|
| Expansion, ordinary recession or manageable AI adjustment | Dominant scenario. Current earnings, employment and financing access provide substantial resilience. | Corrections and investment cuts can occur without the sustained, index-wide destruction needed for a qualifying close. This mainly supports NO. |
| Sustained AI and technology investment bust, without a generalized crisis | Smaller but material scenario, and the largest contributor to YES risk. Spending commitments and financing dependence make a prolonged adjustment credible. | Weak customer returns reduce investment and supplier earnings while investors pay less for those earnings. A sufficiently persistent combination could cross the threshold, though a major technology downturn need not do so. |
| Severe generalized financial or economic crisis | Least likely scenario, but the most dangerous. Broad crisis conditions were not evident at the cutoff, although AI-related financial links create a transmission channel. | Earnings deterioration, restricted financing and falling valuations reinforce one another. An AI bust that becomes a generalized crisis belongs here, rather than being counted twice. Even this scenario does not guarantee a qualifying decline. |
Implications
- NO is not a forecast of benign markets. A substantial bear market, recession or disappointing AI investment cycle could occur without satisfying this unusually deep threshold.
- The relevant stress is correlated. For risk planning, the concern is simultaneous pressure on technology earnings, infrastructure investment and customer financing—not merely losses in isolated speculative companies.
- Cash collection matters more than headline commitments alone. Strong contracted demand is more reassuring when supported by independently funded customer payments. Conversely, guarantee ceilings and multiyear commitments should not be treated as expected losses or immediate cash obligations.
- The price path matters, not just the end-2030 level. A temporary qualifying close counts. A later speculative rise could also raise the reference peak before a subsequent collapse.
Signs to watch
| Timing or development | What would change the assessment |
|---|---|
| November 10, 2026: scheduled Siblis Nasdaq-100 valuation update | Valuations becoming more demanding relative to delivered earnings would make YES more likely; earnings catching up with prices would make it less likely. This is a secondary Nasdaq-100 indicator, not a direct Composite valuation measure. |
| Subsequent quarterly reporting and 2027 investment plans | Cash generation catching up with infrastructure spending, supported by customer collections, would make YES less likely. Underutilization-driven investment cuts, weaker collections or broad earnings downgrades would make it more likely. A convincing reconciliation of customer cash receipts with infrastructure commitments would settle a major uncertainty and make the forecast worth rerunning. |
| An unexpected funding interruption, customer default or guarantee impairment | Evidence that distress is spreading between AI customers, suppliers and lenders would make YES more likely. Reduced financing dependence and demonstrably smaller counterparty exposures would make it less likely. A major event of this kind would warrant rerunning the forecast. |
| Fiscal 2029: expected first phase of Nvidia’s guarantee | The effective exposure depends on construction and lease commencement. Activation alongside weakening customer finances would make YES more likely; reduced exposure or stronger supporting cash flows would make it less likely. Material clarification of the effective terms would justify rerunning the forecast. |
| Macroeconomic, policy and market developments through December 31, 2030 | A persistent closing drawdown accompanied by broad profit downgrades, widening credit spreads and inflation that constrains policy support would make YES more likely. Resilient earnings, improving cash flow, easing inflation and open credit markets would make it less likely. A new record alone is not decisive: it raises the reference peak, while the earnings and financing behind it determine whether vulnerability is increasing. |
Research and methodology
The forecast combines historical drawdown analysis with current company disclosures and economic conditions.
Historical analysis. The research assembled 14,058 daily closes from February 5, 1971 through November 6, 2026, using Nasdaq-sourced FRED history and the latest Nasdaq close. It calculated each day’s decline from the cumulative closing record, identified independent peak-to-recovery episodes, and examined historical windows comparable to the remaining horizon.
It also replayed complete historical return sequences and resampled consecutive two- and three-year blocks to construct possible future paths. Each path checked daily closes and updated its reference peak when a new high occurred. Longer blocks received greater emphasis because short blocks break apart the persistent multiyear deterioration central to this question. These calculations informed the scenarios; they were not treated as independent additional examples of extreme crashes.
The sources with the greatest bearing were:
- Market data and resolution definitions: Nasdaq’s Composite index page and FRED’s NASDAQCOM series supplied the current position and historical closing record.
- Contractual and financing exposures: SEC filings from Amazon, Nvidia and Meta, together with the IMF’s AI-financing assessment, informed the contagion scenario.
- Valuation evidence: Nasdaq’s October 30 dashboard provided the verified forward-earnings comparison; Siblis supplied supplementary valuation context and its scheduled update date.
- Economic and credit conditions: the November 6 employment release, October 28 Federal Reserve statement and FRED high-yield credit-spread series established the starting economic and financing conditions. Credit spreads measure the extra yield required for riskier corporate debt.
Limitations
- The historical tail is poorly measured. There is only one independent threshold-crossing episode. Thousands of overlapping windows and simulated paths cannot turn it into a large independent sample, and future market composition and crisis mechanisms may differ.
- A consistent Composite-wide valuation series could not be verified. The research did not obtain reliable, cutoff-consistent forward and cycle-normalized valuations for the entire Composite. Nasdaq-100 figures cannot simply substitute for them.
- The AI financing map remains incomplete. Public disclosures do not fully reveal counterparties, cancellation rights, collateral or potential lender losses; parts of Nvidia’s guarantee economics are redacted. The most valuable missing evidence is a reconciliation of independently funded end-customer cash receipts against infrastructure commitments and financing-linked demand.
The 10.6% forecast should therefore be understood as a judgment under substantial uncertainty, not as precision established by a large sample of comparable events.
Forecast history
Hover a data point to trace its series, or click to view the forecast generated at that time.
Signed forecast receipt
Signed Nov 7, 2026, 1:52 AM with ed25519 key preseen-prod-ed25519-20260523 and externally timestamped Nov 7, 2026, 1:52 AM.
sha256:96a74018167290...1fabd74e4aWant forecasts like this?
Preseen researches your questions and answers them with calibrated, source-backed forecasts.