Will NASDAQ fall at least 65% from its historical peak at any time before the end of 2030?
Latest forecast
Generated Nov 12, 2026, 11:24 PM
Write-up
Executive Summary
The probability of YES is 11.0%. The Nasdaq Composite is unlikely to record an official closing value at least 65% below its running all-time closing high by December 31, 2030, but the risk is meaningful enough to warrant contingency planning.
The threshold is substantially more severe than an ordinary technology bear market. On November 12, 2026, the index was 8.09% below its record. Without an intervening new high, it would need to fall another 61.92% from that close to qualify.
AI-related investment raises the risk through large spending commitments, financing dependence and concentrated exposures. Substantial operating cash generation and currently orderly credit markets are the strongest counterweights. A single qualifying daily close would settle YES, even if the index subsequently recovered.
Routine update
The update reinforces the existing concern that an AI investment bust could spread through financing and supplier earnings, rather than identifying a new crisis. The chance of a 65% Nasdaq Composite closing drawdown by end-2030 remains about 11%, up just 0.4 percentage points; avoiding that threshold remains the clear expectation.
Key drivers
Historical precedent is rare—and concentrated in one cycle
The daily closing history since 1971 contains only one distinct record-to-recovery cycle that crossed the threshold: the dot-com collapse, which ultimately reached a 77.93% closing drawdown. Even the 1973–74 decline stopped short, at 59.90%.
The financial crisis also produced qualifying closes against the still-unrecovered March 2000 record. Those observations matter under this question’s rules, but they are not a second independent collapse from a fresh all-time high. Repeatedly counting the same prolonged episode would overstate the historical evidence.
AI investment is creating genuine funding pressure
For July–September 2026, Microsoft, Alphabet, Amazon and Meta together reported $201.923 billion of cash capital expenditure—cash spent on property and equipment—against $175.498 billion of operating cash flow.
This standardized comparison shows that investment is no longer entirely covered by current operating cash generation across that group. It does not establish insolvency: the spending is company-wide, not exclusively AI-related; the companies retain financial resources; and investment can be moderated. Their cash balances also cannot be treated as a single pooled account.
Financing links can amplify an investment disappointment
The IMF and Kansas City Fed research identifies connections among equity investments, customer financing, infrastructure debt, purchase commitments and guarantees. If a major AI customer encounters difficulty, suppliers can lose expected revenue while also suffering investment losses or financing exposure.
Concentration strengthens this channel. Dated Nasdaq factsheet observations place technology above 60% of the Composite and its ten largest securities above half its weight. A qualifying collapse would nevertheless require severe losses across major constituents, substantial spillovers, or greater concentration at a future peak—not merely failures among smaller speculative companies.
Real businesses and calm credit markets restrain the risk
The largest infrastructure buyers generate substantial revenue, operating profit and cash flow. Microsoft’s latest quarter, for example, showed strong Azure growth alongside substantial operating income. This is not a simple replay of an investment boom dominated by businesses without established revenue.
Public credit markets also remained orderly. The latest high-yield credit spread—the extra yield on riskier corporate debt relative to safer benchmark debt—was 330 basis points on November 11. Broad financial conditions were looser than average in the week ending November 6.
Lower valuation multiples help, but do not provide a floor
Research found evidence of declining forward price-to-earnings ratios: share prices relative to expected future earnings. That reduces exposure to a crash driven solely by investors paying less for unchanged earnings.
However, much of the apparent improvement relies on higher earnings forecasts. Those forecasts could fall alongside demand, utilization and margins. The available valuation measures also concern the S&P 500 or Nasdaq-100, not a verified contemporaneous Nasdaq Composite valuation.
Scenarios
| Path | How it develops and resolves | Why it carries this weight |
|---|---|---|
| AI growth validates investment, or spending adjusts without a systemic bust | Commissioned capacity produces durable cash returns, financing dependence declines, or investment slows without a broad revenue collapse. Corrections may occur, but the Composite does not breach the threshold: NO. | This receives the greatest weight because major incumbents remain substantial cash-generating businesses and current credit conditions do not indicate an active systemic crisis. |
| AI investment bust becomes an earnings-and-financing feedback loop | Disappointing customer returns weaken cash collections and infrastructure utilization. Procurement cuts hurt suppliers, financing tightens, and losses spread across major constituents. A sufficiently deep, sustained decline produces YES. | This is the main adverse pathway. Large commitments and interconnected exposures make it credible, but it requires several reinforcing failures—not simply slower AI growth or lower share valuations. |
| A broader economic or financial shock compounds technology weakness | An earnings recession, persistently high inflation-adjusted interest rates, or forced selling by leveraged investors coincides with technology-sector stress. YES becomes possible if the combined damage is sufficiently severe. | This remains a secondary adverse pathway. The research identified amplification risks, but current credit and financial-condition indicators argue against treating such a crisis as already underway. |
Implications
- Treat a qualifying collapse as a contingency, not the central planning assumption. The forecast supports testing resilience to an extreme downside outcome without making that outcome the default expectation.
- NO is not a bullish return forecast. A substantial bear market or disappointing technology returns could occur without reaching this unusually demanding threshold.
- Look beyond direct equity exposure. AI customers, suppliers, infrastructure financiers and guarantors can share exposure to the same underlying spending. Apparently different investments or business relationships may therefore deteriorate together.
- Cash returns matter more than growth headlines alone. Strong adoption can coexist with excessive capacity, falling margins or weak investment returns. Conversely, spending moderation need not imply a catastrophic bust.
Signs to watch
The research did not verify exact dates for future earnings releases or policy meetings, so no such calendar is assumed here. The monitoring horizon runs through December 31, 2030.
| Timing or development | What to watch | Direction for the forecast |
|---|---|---|
| Subsequent quarterly disclosures | Cash generation catches up with investment while growth remains durable; alternatively, spending continues to outpace cash generation alongside rising financing needs. | Improving cash coverage makes YES less likely; persistent deterioration makes it more likely. |
| Customer or financing announcements, potentially without warning | Major capacity-contract cancellations, weaker customer collections, refinancing difficulties or guarantee calls spread across AI counterparties. | Makes YES more likely. If these appear together across major suppliers and customers, the forecast is worth rerunning because a central transmission risk would be becoming observable. |
| Daily and weekly market indicators | Credit spreads widen, financing conditions tighten and selling broadens beyond the largest technology shares. | Makes YES more likely. Continued orderly credit alongside broader earnings growth makes it less likely. |
| Capacity commissioning and operating disclosures | New infrastructure delivers durable cash returns, utilization remains healthy and dependence on external financing declines. | Makes YES less likely. Sustained evidence of this would justify rerunning the forecast because it would resolve a major uncertainty about the build-out. |
| Macroeconomic and policy developments | High inflation-adjusted interest rates persist while earnings forecasts weaken, or forced selling emerges among leveraged investors. | Makes YES more likely, particularly if it coincides with AI-sector cash-flow stress. |
| Official daily closes through December 31, 2030 | Track each close against the highest official closing value reached up to that date. New records reset the reference peak. | A qualifying close settles YES, regardless of any later recovery. Without one by the deadline, the question settles NO. |
Research and methodology
The forecast combines original historical calculations with a judgmental assessment of current investment and financing conditions. Its principal sources were:
- Nasdaq official index history and Nasdaq-sourced FRED data. The historical dataset contained 14,062 nonmissing daily closes from February 5, 1971 through November 12, 2026. FRED supplied the history through November 11; the official November 12 close completed the sample. Missing observations were dropped rather than filled.
- IMF AI-ecosystem assessment, October financial-stability analysis and Kansas City Fed financing research. These established plausible channels connecting investment disappointment to customer, supplier and financing losses.
- Nasdaq Composite factsheet. This informed the concentration assessment, with explicit caution about inconsistent document vintages.
- High-yield credit spreads and Chicago Fed financial conditions. These provided evidence against an already-developed broad funding crisis.
- FactSet and Citadel Securities. These informed the direction of valuation changes, not an exact Composite valuation.
The historical work identified complete record-to-recovery drawdown cycles, examined forward windows matching the remaining horizon, replayed intact historical return paths and simulated paths by resampling consecutive stretches of returns. Longer stretches received greater attention because short stretches can interrupt the prolonged deterioration needed for this event. Every path updated its reference peak after new closing highs.
These methods share the same underlying history and were not treated as independent confirmations. Current-condition adjustments were judgmental rather than statistically estimated AI-risk coefficients. The observation window was interpreted prospectively from the client’s July 21, 2026 request; earlier peaks remain valid references, but earlier crashes do not themselves resolve the question.
Limitations
The greatest limitation is the single independent historical qualifying cycle. Overlapping windows and repeated simulations cannot turn that episode into a large sample, and numerical precision should not be mistaken for certainty about future market behavior.
Research did not establish a reliable contemporaneous Composite forward price-to-earnings ratio, exact November constituent weights, or a complete map of AI-customer financing and supplier backlog exposure. Factsheet vintages and some valuation exhibits were inconsistent.
Better disclosure of contract cancellation rights, guarantees, capacity utilization and end-customer cash returns would materially improve the assessment of whether today’s commitments are sustainable.
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 12, 2026, 11:24 PM with ed25519 key preseen-prod-ed25519-20260523 and externally timestamped Nov 12, 2026, 11:24 PM.
sha256:013513bb063cee...05d116be34Want forecasts like this?
Preseen researches your questions and answers them with calibrated, source-backed forecasts.