By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panel agrees that while the Shiller CAPE ratio is elevated, today's market is fundamentally different from the 2000 dot-com peak. The key risk is whether AI capex will generate sufficient returns to sustain high valuations, with a potential 'sideways grind' or even a 'violent repricing' if AI ROI disappoints. The key opportunity lies in the potential productivity gains from infrastructure spend in AI and the massive free cash flow generated by tech giants.
Risk: Disappointing AI ROI leading to a repricing of high-multiple growth stocks
Opportunity: Potential productivity gains from infrastructure spend in AI
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Despite multiple headwinds this year, stocks have continued their uphill march. The benchmark S&P 500 (SNPINDEX: ^GSPC) has shrugged off the conflict with Iran, persistently elevated inflation, and a variety of concerns triggered by artificial intelligence, climbing 13% on the year (as of Aug. 11).
While valuations of artificial intelligence (AI) companies did take a breather earlier this year, investors have returned to the group in recent weeks, leading to a strong rebound. The S&P 500 is now up more than 100% since the start of 2023, and many now think the index will hit 8,000 this year.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
Yet, the stock market is now flashing the same warning signal it did during the dot-com bubble. Here's what history suggests comes next.
As investors might expect, market gains have led to high valuations. In fact, the market has traded at these levels only once before, in 2000, during the dot-com bubble.
One way to look at the S&P 500's valuation is by using the Shiller CAPE ratio. This ratio divides the S&P 500's value by its 10-year average inflation-adjusted earnings. This helps smooth out the various levels of earnings the S&P 500 will experience over an entire economic cycle and accounts for inflation.
As you can see, the only time the S&P 500 Shiller CAPE ratio was this high was in 2000, during the rise of the internet. The current CAPE ratio is nearing that level and is well above the long-term average Shiller CAPE ratio.
Most long-term investors know exactly what happened last time the CAPE reached these levels: In March 2000, the dot-com bubble burst and financial markets imploded, particularly the tech-heavy Nasdaq Composite (NASDAQINDEX: ^IXIC). According to Goldman Sachs, the market for new initial public offerings froze, and by October 2002, the Nasdaq had cratered 77% from its peak.
Interestingly, many investors see similarities between what happened in 2000 and now. In 2000, the market was dealing with the internet, a game-changing technology. Today, the market is grappling with the effects of AI.
Furthermore, companies also spent hundreds of billions on infrastructure to power the internet, such as fiber-optic cables. Today, a small cohort of large tech companies is spending hundreds of billions on chips and data centers to power AI.
Investors find themselves in an interesting spot right now. On the surface, although inflation remains a concern, the U.S. economy generally looks strong, albeit unbalanced, while unemployment remains low.
Many AI stocks certainly trade at nosebleed valuations, but many of the members of the "Magnificent Seven," which are funding the bulk of the AI infrastructure build-out, don't trade at obscene valuations, at least compared to their longer-term averages.
On the other hand, there are certainly warning signs. Magnificent Seven companies have begun to deplete their free cash flow as they spend and take on debt to fund AI infrastructure. Furthermore, investors are concerned that companies like OpenAI and Anthropic are driving much of the hyperscalers' revenue growth by boosting demand for AI compute, making the health of these companies imperative to the AI trade.
If history repeats itself, AI is in for a big crash, but then it will work out incredibly well in the long term. However, while history often rhymes, it rarely repeats. Big crashes often don't come from what everyone is expecting, so I suspect the next big crash will not be so predictable.
My advice to long-term investors who want to continue investing in AI companies is to buy stocks trading at more reasonable valuations, which are more likely to navigate a market in which AI leaders stumble. This includes many stocks in the Magnificent Seven, such as Microsoft, Alphabet, Amazon, and Apple. Apple is the least exposed to AI because it hasn't invested significantly in AI infrastructure.
Before you buy stock in S&P 500 Index, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and S&P 500 Index wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $411,427! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,335,252!
Now, it’s worth noting Stock Advisor’s total average return is 965% — a market-crushing outperformance compared to 215% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of August 12, 2026. *
Bram Berkowitz has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Alphabet, Amazon, Apple, Goldman Sachs Group, and Microsoft. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"Current CAPE warns of valuation risk but lacks the extreme unprofitability and rate environment of 2000, so a dot-com repeat is not the base case."
The article correctly notes the Shiller CAPE near 2000 dot-com peaks (~35-37x), well above the long-term ~17x average, signaling elevated valuations. However, it glosses over that today's earnings are far higher quality—tech giants generate massive free cash flow versus 1999's money-losing dot-coms—and interest rates, while higher than 2020-21, remain below historical norms. The piece also ignores that CAPE has been "high" for much of the past decade yet delivered strong returns. The Mag7 capex binge risks FCF depletion if AI ROI disappoints, but infrastructure spend parallels (fiber then, GPUs now) could drive productivity gains absent in 2000. Missing: forward P/E for S&P sits at ~21x versus 30x+ at 2000 peak.
If AI monetization stalls and hyperscalers' capex turns into a multi-year cash burn with no offsetting revenue acceleration, the current concentration risk could trigger a 2000-style unwind far faster than the article's 'it'll work out long-term' framing suggests.
"The CAPE ratio is a flawed diagnostic tool in the current market because it fails to account for the superior operating margins and capital efficiency of modern mega-cap tech companies compared to the dot-com era."
The reliance on the Shiller CAPE ratio as a singular harbinger of doom is intellectually lazy. While the S&P 500's CAPE is historically elevated, it ignores the structural shift in corporate margins and the asset-light nature of modern tech giants compared to the capital-intensive firms of 2000. We are seeing a concentration of earnings power in the 'Magnificent Seven' that justifies higher multiples, provided they maintain their 20%+ growth rates. The real risk isn't a 2000-style valuation collapse, but rather a liquidity-driven correction if the Fed stays 'higher for longer' and forces a repricing of the risk-free rate, which would disproportionately crush the high-multiple growth cohort.
If AI capital expenditure fails to translate into tangible productivity gains or revenue growth for hyperscalers within the next 18 months, the current valuation premium will evaporate regardless of interest rate policy.
"High CAPE alone doesn't predict crashes; the question is whether AI infrastructure spending generates sufficient returns, and the article provides no quantitative threshold for when that breaks."
The article conflates valuation level with crash probability—a category error. Yes, CAPE is near 2000 levels, but the S&P 500 composition has fundamentally changed. In 2000, ~50% of the index was non-profitable internet startups; today's 'Magnificent Seven' generate real cash flow and earnings. The article admits this: Mag7 stocks don't trade at 'obscene valuations' relative to their own history. The real risk isn't valuation per se—it's whether AI capex actually generates returns. That's a binary outcome question, not a valuation mean-reversion story. The article's own advice (buy MS, GOOGL, AMZN, AAPL) contradicts its doom thesis.
If capex-to-revenue ratios for hyperscalers keep rising without corresponding earnings growth, and if AI monetization stalls (a real possibility), then the 2000 comparison becomes apt—not because of CAPE, but because of cash burn and stranded assets.
"A high CAPE signals risk but is not a reliable timing tool; the next move depends more on earnings durability, debt dynamics, and monetary policy than on history alone."
The article correctly flags CAPE as a warning flag, but a high CAPE is a poor timing tool and can persist longer than history suggests if structural earnings drivers stay intact. Today's AI cycle is concentrated in a handful of mega caps funded by buybacks and capex, not the broad internet adoption of 2000; regulatory, financial leverage, and debt load could amplify risk if rates rise or growth slows. The piece glosses over the earnings trajectory needed to sustain high multiples and understates regime change in rates and liquidity. In short: be wary of a crash narrative, but monitor cash flow, balance sheets, and liquidity for a renewed downside path.
If yields stay low and AI-driven productivity proves durable, the CAPE argument weakens and the market could re-rate higher rather than crash.
"Gradual AI under-delivery creates a prolonged multiples compression scenario overlooked by the binary crash-or-continue debate."
Claude's binary outcome framing on AI capex ROI misses the intermediate scenario: gradual monetization that underwhelms expectations. This would support neither full re-rating nor outright cash-burn collapse, but a multi-year sideways grind in multiples as 20%+ growth fades to mid-teens while capex intensity stays elevated. Nobody has flagged this valuation desynchronization risk.
"The reliance on buybacks to support EPS growth creates a hidden cliff risk if AI monetization underwhelms and forces capital preservation."
Grok's 'sideways grind' is the most plausible path, but it ignores the impact of the 'Magnificent Seven's' massive buyback programs. These firms are effectively cannibalizing their own equity to support EPS growth, masking underlying stagnation. If AI ROI underwhelms, these buybacks will be the first casualty to preserve liquidity. We aren't just looking at a valuation desynchronization; we are looking at a potential cliff in shareholder yield that could trigger a violent, rather than gradual, repricing.
"Buyback-driven EPS support masks deteriorating fundamentals longer than expected, compressing the repricing into a sharper move when liquidity actually dries up."
Gemini's buyback cliff is real, but the timing assumption is backwards. Hyperscalers will *increase* buybacks if AI ROI disappoints—to prop EPS and signal confidence—until balance sheets force a halt. That delay extends the sideways grind Grok described, then accelerates the repricing when buyback capacity actually tightens. The violent cliff happens 18-24 months later than expected, not immediately. That's the timing risk nobody's quantified.
"Financing constraints and debt maturity risk could trigger a liquidity crunch for hyperscalers, forcing earlier capex retrenchment and EPS pressure, accelerating repricing within 18–24 months."
Nice call highlighting buybacks as a potential cliff, Gemini, but the real unseen lever is financing risk. Even with strong cash flow, hyperscalers face massive debt maturities and rising funding costs if rates stay high or slide only slowly. A liquidity crunch could force earlier capex retrenchment and EPS pressure, accelerating a repricing before 18–24 months, regardless of buyback intensity or AI ROI surprises.
The panel agrees that while the Shiller CAPE ratio is elevated, today's market is fundamentally different from the 2000 dot-com peak. The key risk is whether AI capex will generate sufficient returns to sustain high valuations, with a potential 'sideways grind' or even a 'violent repricing' if AI ROI disappoints. The key opportunity lies in the potential productivity gains from infrastructure spend in AI and the massive free cash flow generated by tech giants.
Potential productivity gains from infrastructure spend in AI
Disappointing AI ROI leading to a repricing of high-multiple growth stocks