AI Can Win and AI Stocks Can Still Crash: The Valuation Risk Traders Are Missing
There is an assumption buried inside much of the artificial intelligence trade that I think traders need to reconsider.
It goes something like this:
If AI really does transform the economy, the stocks leading the AI revolution should continue moving higher.
That sounds logical.
But it isn’t necessarily true.
A group of European Central Bank economists has raised a fascinating possibility: a significant AI stock market correction may eventually occur even if artificial intelligence lives up to the extraordinary expectations investors currently have for it.
In other words, Nvidia does not have to fail for NVDA to fall.
Microsoft does not have to lose money on artificial intelligence for MSFT to correct.
AI does not have to be a bubble for technology valuations to come down.
The technology can be enormously successful while the stocks associated with it still experience a painful repricing.
That distinction may be one of the most important things investors and traders understand about the next stage of the AI boom.
The ECB Is Not Saying AI Is a Fraud
The ECB economists’ argument is more sophisticated than the familiar comparison between today’s artificial intelligence boom and the dot-com bubble.
In fact, they specifically argue that a correction in technology valuations should not automatically be interpreted as evidence that investors were irrational.
Looking at previous technological revolutions, the researchers conclude that a correction in today’s elevated valuations is likely even if those valuations were rational when investors initially established them. ([European Central Bank](https://www.ecb.europa.eu/press/blog/date/2026/html/ecb.blog20260817~754a8a4418.en.html?utm_source=chatgpt.com))
That is an important distinction.
Artificial intelligence can be genuinely transformative.
AI companies can generate tremendous revenue.
Productivity can improve.
Entire industries can be reorganized around the technology.
And some of today’s leading AI stocks can still decline sharply.
To understand why, we need to separate the value of the technology from the valuation of the stock.
A Great Company Can Still Be an Expensive Stock
This is one of the most basic principles of investing, but it is remarkably easy to forget during powerful bull markets.
A company can perform spectacularly while its stock performs poorly.
Why?
Because the price investors are willing to pay for those future earnings can change.
Imagine a company earns $10 per share and investors are willing to pay 40 times earnings for it.
The stock would theoretically trade at $400.
Now suppose earnings grow 20% to $12 per share.
That is fantastic operational performance.
But if investors decide the appropriate valuation is now only 25 times earnings, the stock is worth:
$12 × 25 = $300.
Earnings increased 20%.
The stock fell 25%.
Nothing went wrong with the business.
The multiple changed.
That is the valuation risk facing the artificial intelligence trade.
The Option Value of Nvidia
The ECB economists use an idea that is particularly useful for understanding what has happened to stocks such as Nvidia (NVDA).
Think about Nvidia several years ago.
Investors knew artificial intelligence might become enormously important.
But nobody knew exactly how important.
There was a wide range of possible outcomes.
At one extreme, AI adoption could have disappointed and Nvidia’s enormous investment in accelerated computing might never have produced the revenue investors expected.
At the other extreme, Nvidia could become the central computing platform for one of the largest technological transformations in history.
That extreme upside potential has value.
Economists describe it as option value.
The ECB’s researchers argue that uncertainty surrounding an emerging breakthrough can actually justify very high valuations for pioneering companies because investors are purchasing exposure to potentially extraordinary outcomes. ([European Central Bank](https://www.ecb.europa.eu/press/blog/date/2026/html/ecb.blog20260817~754a8a4418.en.html?utm_source=chatgpt.com))
That helps explain why investors may rationally pay exceptionally high multiples for early technological leaders.
But Something Changes When the Technology Succeeds
Here is the counterintuitive part.
Suppose artificial intelligence succeeds.
Not partially.
Completely.
AI spreads through banking, healthcare, manufacturing, logistics, advertising, software, transportation, defense, education and virtually every other major industry.
At that point, AI is no longer simply an opportunity belonging to a handful of companies.
It becomes part of the economy itself.
And according to the ECB researchers, that changes the nature of the risk.
During the early stage, the uncertainty is concentrated in individual companies.
Will Nvidia win?
Will Microsoft win?
Will OpenAI succeed?
Will another architecture replace today’s technology?
But once AI becomes deeply embedded throughout the economy, a disruption involving the technology could affect virtually everyone.
Company-specific risk becomes increasingly systemic risk.
AI Success Can Actually Increase the Risk Premium
This is the heart of the ECB argument.
As artificial intelligence becomes more important to the entire economy, investors may require greater compensation for owning stocks exposed to that economy-wide risk.
Economists call that compensation the equity risk premium.
And when the required return investors demand from stocks increases, valuations generally decline.
That creates a seemingly strange sequence:
AI succeeds → AI spreads throughout the economy → systemic exposure to AI increases → investors demand a larger risk premium → valuation multiples compress.
The technology succeeds.
Profits can rise.
Productivity can increase.
Yet stock prices can still fall because investors are no longer willing to pay the same multiple for those earnings.
We Have Seen This Movie Before
The ECB economists compared artificial intelligence with several previous technological revolutions, including:
- Railroads in the 19th century
- Electrification
- Radio in the 1920s
- The internet during the dot-com era
These technologies were not imaginary.
They changed the world.
Railroads transformed transportation and commerce.
Electricity transformed almost every industry.
Radio transformed mass communication.
The internet became one of the foundational technologies of the modern economy.
And yet stocks associated with those innovations experienced enormous booms and painful corrections along the way.
The ECB researchers note that transformative technologies have historically attracted investment and driven valuations sharply higher before those valuations eventually declined. ([European Central Bank](https://www.ecb.europa.eu/press/blog/date/2026/html/ecb.blog20260817~754a8a4418.en.html?utm_source=chatgpt.com))
The lesson isn’t that new technologies fail.
The lesson is that technological success does not guarantee continuously rising stock valuations.
The Internet Is Probably the Best Comparison
Think about the internet in 1999.
Investors were absolutely correct about the technology.
The internet did transform business.
It transformed communication.
It transformed shopping.
It transformed entertainment.
It transformed advertising.
It created some of the largest companies in history.
The technology wasn’t the mistake.
The mistake was assuming that every price investors were willing to pay for internet exposure could be justified by the eventual economics.
That is the distinction traders need to remember today.
Artificial intelligence may prove every bit as transformative as its strongest advocates believe.
That does not tell us what multiple investors should pay for AI earnings today.
Why Nvidia Can Keep Growing and NVDA Can Still Correct
Let’s bring this directly back to Nvidia.
Nvidia could continue selling extraordinary numbers of GPUs.
Data-center revenue could continue growing.
New generations of Nvidia processors could remain dominant.
OpenAI, Microsoft, Meta, Amazon, Google and other customers could continue building enormous AI infrastructure projects.
And NVDA could still experience a substantial correction.
That would not necessarily invalidate the Nvidia story.
It could simply mean investors were willing to pay less for each dollar of future Nvidia earnings.
This is why traders should watch both:
earnings growth
and
valuation multiples.
They are not the same thing.
The Bond Market Makes This Even More Important
This argument becomes even more compelling when we connect it to what is currently happening in the bond market.
Long-term government yields have been rising as investors confront persistent inflation, enormous fiscal deficits, heavy government borrowing and rapidly increasing corporate debt issuance associated with the artificial intelligence infrastructure boom.
This matters because interest rates influence the discount rate investors use to value future earnings.
Higher long-term yields mean future corporate profits are worth less in present-value terms.
That disproportionately affects companies whose valuations depend on earnings expected many years into the future.
In other words, many technology companies are effectively long-duration assets.
The combination of rising Treasury yields and elevated AI valuations therefore creates a particularly important risk.
The AI Boom May Be Creating Its Own Headwind
This brings together the three major themes we have been following recently at TraderInsight.
First, artificial intelligence companies require enormous amounts of infrastructure.
Second, companies such as Nvidia are beginning to help finance that infrastructure.
Third, those investments are contributing to unprecedented demand for capital at the same time governments are issuing enormous quantities of debt.
That creates a fascinating feedback loop:
AI demand grows → AI infrastructure spending grows → borrowing grows → long-term capital becomes more expensive → discount rates rise → AI valuation multiples come under pressure.
That does not mean AI growth stops.
In fact, the opposite may be happening.
The industry may be growing so rapidly that financing that growth itself begins affecting asset valuations.
Why Europe Is Worried About American Technology Stocks
The ECB’s concern goes well beyond whether NVDA or QQQ experiences a correction.
American technology stocks have become deeply embedded in European household and institutional portfolios.
Euro-area investors have built substantial exposure to U.S. equities through investment funds, pension products and insurers. The ECB has repeatedly highlighted the increasing importance of U.S. technology and AI-related stocks to European financial portfolios. ([European Central Bank](https://www.ecb.europa.eu/press/financial-stability-publications/fsr/html/ecb.fsr202605~50566915a7.en.html?utm_source=chatgpt.com))
That means a sharp technology selloff in the United States would not remain neatly contained inside American brokerage accounts.
European investment funds could experience losses.
Pension assets could decline.
Insurers could see investment portfolios fall.
Risk appetite could deteriorate.
And because U.S. and European equity markets often move together during periods of stress, falling American technology stocks could quickly become a broader global financial event.
Market Concentration Adds Another Layer of Risk
There is also the issue of concentration.
The Magnificent Seven—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla—have become an extraordinarily important component of U.S. equity-market performance.
ECB analysis earlier this year noted that the Magnificent Seven accounted for roughly 40% of the S&P 500’s market capitalization. ([European Central Bank](https://www.ecb.europa.eu/press/key/date/2026/html/ecb.sp260323_1~1e06784a89.en.html?utm_source=chatgpt.com))
That creates a structural vulnerability.
When a small group of very large companies drives a disproportionate amount of index performance, weakness in those stocks can pull down an otherwise healthy market.
Index investors may believe they own a broadly diversified portfolio while actually carrying substantial exposure to a relatively small collection of mega-cap technology companies.
That concentration becomes particularly important if institutional investors begin reducing exposure simultaneously.
Stocks Traders Should Watch
If this valuation-reset thesis begins playing out, I would concentrate on the stocks carrying the greatest influence over AI sentiment and major indexes.
- Nvidia (NVDA)
- Microsoft (MSFT)
- Meta Platforms (META)
- Amazon (AMZN)
- Alphabet (GOOGL)
- Broadcom (AVGO)
- Oracle (ORCL)
- Apple (AAPL)
The important signal would not simply be whether these stocks decline.
Corrections happen all the time.
The more meaningful signal would be a persistent change in how investors respond to good news.
Watch How Stocks React to Good News
This is one of my favorite ways to judge changes in institutional sentiment.
During a strong bull market, good news tends to produce strong upside reactions.
Earnings beat expectations.
The stock gaps higher and holds the move.
A company announces another AI partnership.
Buyers chase it.
Revenue guidance increases.
The stock expands its valuation multiple.
But mature market cycles often begin behaving differently.
A company reports excellent earnings—and the stock barely moves.
Revenue beats expectations—and sellers appear.
Another massive AI spending announcement is made—and investors begin asking about return on capital instead of celebrating the size of the investment.
That change in reaction is important.
When good news stops producing good price action, pay attention.
Nvidia May Be the Most Important Tell
NVDA remains one of the best sentiment indicators for the entire artificial intelligence trade.
There are several things I would watch closely:
- NVDA relative strength versus QQQ
- Price reaction following earnings beats
- Institutional response to large AI infrastructure commitments
- Whether previous support levels continue attracting buyers
- Volume accompanying breakdowns
- Whether rallies recover quickly or begin failing below prior highs
If Nvidia continues producing spectacular operating results but the stock begins persistently underperforming, that could be an early indication that the market is changing the multiple it is willing to assign to AI growth.
QQQ May Matter More Than Any Single Stock
The Invesco QQQ Trust (QQQ) gives traders a broader view of mega-cap growth appetite.
In a normal pullback, individual leaders may correct while QQQ holds major support and institutional buyers rotate among technology stocks.
A more meaningful valuation reset would look different.
We would expect to see:
- Multiple mega-cap leaders weakening simultaneously
- QQQ losing important support
- Failed rallies rather than immediate dip buying
- Declining market breadth
- Persistent weakness even following favorable earnings
That would suggest investors are not simply rotating between technology leaders.
They may be reducing the valuation premium assigned to the entire group.
Don’t Ignore TLT and Treasury Yields
For traders following AI stocks, I would continue keeping the bond market on the screen.
The iShares 20+ Year Treasury Bond ETF (TLT) provides a convenient way to monitor long-duration Treasury prices.
When TLT falls, long-term yields are generally rising.
That means one particularly dangerous combination would be:
- TLT breaking support
- 30-year Treasury yields making new highs
- QQQ losing support
- NVDA underperforming QQQ
- Other Magnificent Seven stocks failing to respond to positive news
That combination would suggest that rising discount rates and valuation compression are beginning to reinforce one another.
There Is Also a Bullish Scenario
None of this means traders should automatically become bearish on artificial intelligence.
The most constructive environment would be one in which earnings continue growing while valuation excess gradually comes out of the market.
A correction can actually create healthier conditions.
If AI stocks decline while corporate fundamentals remain strong, valuations become more attractive.
Companies with genuine competitive advantages can then separate themselves from companies that simply benefited from AI enthusiasm.
That would create exactly the type of environment professional traders want:
dispersion.
The strongest companies begin outperforming the weaker ones.
And stock selection becomes more important than simply owning everything associated with artificial intelligence.
The Next Phase May Be About Return on Investment
For several years, investors rewarded companies for spending more on AI.
$20 billion sounded exciting.
Then $50 billion.
Then $100 billion.
Now infrastructure commitments are measured in hundreds of billions of dollars.
Eventually, investors are going to ask the obvious question:
What return are you earning on all of this money?
That could be the defining question of the next stage of the artificial intelligence investment cycle.
Companies able to convert enormous capital expenditures into equally enormous cash flow may continue commanding premium valuations.
Companies that cannot may experience substantial multiple compression.
The words “AI investment” alone may stop being sufficient.
The Five Charts I Would Watch
If I wanted to monitor the possibility of an AI stock market correction, I would keep five charts readily available:
- NVDA — the primary sentiment barometer for AI infrastructure.
- QQQ — the broader mega-cap technology trend.
- 30-year Treasury yield — the market price of long-duration capital.
- TLT — confirmation from long-term Treasury prices.
- Equal-weight S&P 500 versus the cap-weighted S&P 500 — a useful measure of whether market leadership is broadening or becoming increasingly dependent on mega-cap technology.
I would be particularly interested in divergences among them.
If the S&P 500 remains near highs while NVDA, QQQ and other AI leaders begin weakening beneath the surface, that could provide an early warning before weakness becomes obvious in the headline indexes.
TraderInsight Trading Implications
The ECB’s research creates several practical lessons for traders.
- Do not confuse a great technology with a great entry price. Artificial intelligence can transform the world while individual AI stocks remain overpriced.
- Watch multiples as well as earnings. Rising profits do not guarantee rising stock prices when valuation multiples are contracting.
- Pay attention to reactions to good news. Strong fundamentals combined with weak price reactions can signal that institutional sentiment is changing.
- Monitor Treasury yields. Higher long-term rates increase discount rates and can accelerate valuation compression in high-growth stocks.
- Watch relative strength. The AI stocks that continue outperforming during a broad correction may become the leaders of the next advance.
- Expect increasing differentiation. As the AI boom matures, balance-sheet strength, cash generation and return on invested capital should matter more.
- Remember global contagion. U.S. mega-cap technology stocks have become important holdings throughout the global financial system, so a serious AI selloff may not remain confined to Nasdaq.
The Most Important Distinction
There are really two separate questions traders need to ask about artificial intelligence.
Question one:
Will artificial intelligence transform the global economy?
I think the evidence increasingly suggests that it will.
But then comes question two:
What price should investors pay today for companies expected to benefit from that transformation?
Those are completely different questions.
You can be extraordinarily bullish about artificial intelligence while simultaneously believing some AI stocks are too expensive.
You can believe Nvidia will remain one of the dominant companies of the next decade while recognizing that NVDA can experience 20%, 30% or larger corrections along the way.
And you can believe AI will create enormous economic value without believing today’s valuation multiples will remain permanently elevated.
TraderInsight Bottom Line
The most interesting part of the ECB’s warning is not that artificial intelligence could fail.
It is that artificial intelligence could succeed.
That success itself changes the nature of the investment.
As AI moves from an emerging technology into critical economic infrastructure, the extraordinary uncertainty that helped justify enormous upside valuations begins to disappear.
At the same time, AI becomes increasingly intertwined with the entire economy, potentially increasing systemic risk and the return investors demand for holding equities.
Add rising long-term interest rates, enormous infrastructure financing requirements and historically concentrated equity indexes, and the conditions for an AI stock market correction begin to make considerably more sense.
The takeaway for traders is straightforward.
Don’t ask only whether AI is going to work.
Ask what the market has already priced in.
Ask what investors are willing to pay for those earnings.
Ask whether good news is still producing good price action.
And watch the bond market as closely as you watch Nvidia.
Because the next major correction in AI stocks may not happen because the technology failed.
It may happen precisely while the technology is succeeding.
TraderInsight educational content is provided for informational and educational purposes only and is not investment advice. Trading involves substantial risk, and past performance does not guarantee future results.