Is AI the Next Dot-Com Bubble? 5 Things Index Investors Should Do Now, and 3 Things That Could Ruin You
On Thursday, November 5, 2026, the AI trade finally flinched.
The S&P 500 fell 4.4% in a single session, its worst day since the "Liberation Day" tariff shock of April 2025. The Nasdaq sank 5.5% and closed nearly 12% below the record it set only weeks earlier. Nvidia dropped more than 9%. AMD lost about 17%.
If you own an S&P 500 index fund, you felt it, even though you never bought a single "AI stock."
And maybe a familiar, sick feeling came back. The one from 2000. Or 2008. Or the last time you lost money and swore it would never happen again.
So let's answer the question honestly, with real numbers instead of hype or doom.
The short answer Nobody can tell you whether AI is a bubble that pops next month or the start of a decade-long boom. Today's market shares several dot-com warning signs: record concentration (the 10 biggest companies are about 43% of the S&P 500), valuations near 2000 levels and record borrowing to buy stocks. But it differs in one big way: today's AI leaders are hugely profitable. For index investors, the winning move is not to predict. It is to prepare. Do these 5 things: check how much AI you already own, rebalance, diversify beyond the top 10, keep 2–3 years of spending money safe, and keep investing on autopilot. Avoid these 3: going all-in on AI or leveraged funds, selling everything to "wait for the crash," and borrowing to invest. |
In this guide you'll see what really happened to ordinary investors when the dot-com bubble burst, how today compares, what an AI crash could do to your index fund in dollars, and the simple plan that protects you whichever way this goes.
What's inside:
What the dot-com bubble really did to ordinary investors
Most people remember the dot-com crash as a story about silly websites. That's only half of it.
The other half is about good, real, profitable companies whose share prices simply ran too far ahead of reality. That half is the one that matters for you today.
The silly half: Pets.com
Pets.com sold pet food online. It ran a famous Super Bowl ad with a sock puppet, then went public on the Nasdaq in February 2000 at $11 a share, raising $82.5 million.
The problem? Pet food is heavy and cheap. Shipping it cost a fortune, and the company lost money on sale after sale.
By November 2000 the shares were worth 19 cents. Just 268 days after its stock market debut, Pets.com shut down and liquidated.
The dangerous half: Cisco
Cisco was the opposite of Pets.com. It made the routers and switches the whole internet ran on. Real products. Real customers. Real profits.
On March 27, 2000, Cisco briefly became the most valuable company in the world, worth more than $500 billion. Its shares closed at a split-adjusted record of $80.06.
Then the bubble burst. Cisco lost roughly 90% of its market value. And here is the part every AI investor should tape to the fridge: Cisco's share price did not close above that 2000 high again until December 10, 2025. More than 25 years later.
The business kept growing. The stock didn't recover for a quarter of a century, because investors had paid far too much for it on day one. At its 2000 peak, Cisco traded at more than 200 times its earnings.
The lesson in one line A world-changing technology and a good investment are not the same thing. The internet changed everything. People who paid any price for internet stocks in 2000 still lost for years. |
What happened to the whole market
The tech-heavy Nasdaq Composite peaked at 5,048.62 on March 10, 2000. Over the next 31 months it fell 78%, bottoming at 1,114.11 in October 2002.
It did not close above its 2000 peak until April 23, 2015. Fifteen years.
The broader S&P 500 held up better, but it still fell about 49% from its March 2000 peak to its October 2002 low. Then the 2008 financial crisis hit before investors had fully healed.
The result was the "lost decade." From January 2000 to December 2009, the S&P 500 returned about −1% a year, dividends included. A dollar invested at the start of 2000 was worth roughly 91 cents ten years later.
If that sounds like the worst-case scenario for an index investor, it was. But hold that thought, because the same decade also contains the most hopeful lesson in this article. We'll get to it in Step No. 3.
(If you want the full story of what a 34-year wait looks like, read What If the S&P 500 Does What Japan Did? Lessons From the Nikkei's 34-Year Wait.)
Is AI the next dot-com bubble? 5 ways today looks the same
Here is the uncomfortable evidence, laid out plainly.
No. 1: The index has never been this top-heavy
An S&P 500 fund is supposed to spread your money across 500 companies. But it weights each one by size, so the giants get most of your money.
At the 2000 peak, the 10 biggest companies made up about 27% of the S&P 500. By the end of 2025, the top 10 had reached a record 40.7%, according to RBC Wealth Management. By July 2026, RBC's data put it near 43%.
Even more telling: in 2025 those top 10 companies were about 41% of the index's value but were expected to produce only about 32% of its earnings. In 2015, value and earnings matched at about 19% each. The gap is what investors are paying for future growth that hasn't happened yet.
(For a deeper look at this split, see The Mag 7 Lost the Race to the "Boring" 493 in 2026.)
No. 2: Valuations are back near 2000 levels
The Shiller CAPE ratio compares stock prices with 10 years of inflation-adjusted earnings. Its long-run average is about 17.
At the dot-com peak in December 1999, it hit a record of about 44. In 2026 it has stayed above 40 since May, a level reached only once before: during the dot-com bubble.
High valuations don't tell you when a fall will come. They tell you how little room for error there is.
No. 3: The spending boom is enormous
Building AI takes chips, data centers and electricity, and the biggest tech companies are spending at a pace the world has never seen.
Amazon, Microsoft, Alphabet and Meta spent roughly $410 billion on capital projects in 2025. Their 2026 plans add up to about $725 billion, a jump of roughly 77%, according to a Financial Times tally of company guidance. Bank of America has forecast that these companies' borrowing will hit about $175 billion in 2026, more than six times their prior five-year average.
The dot-com era had its own version: telecom companies laid huge networks of fiber-optic cable for demand that arrived years later than promised. Over-building is a classic bubble ingredient.
No. 4: People are borrowing record amounts to buy stocks
Margin debt is money investors borrow from their broker to buy shares. FINRA's figures show it hit a record of about $1.5 trillion in June 2026, up roughly 49% in a year.
In the late 1990s, margin debt surged into March 2000, right as the market peaked. Borrowed money turns a fall into a forced sale, which turns a fall into a crash.
No. 5: Everyone is talking about it
In November 2025, more than 12,000 news stories mentioned "AI bubble," about as many as the previous ten months combined, according to Bloomberg data. A month earlier, more than half the investors in Bank of America's October 2025 fund manager survey said AI stocks were in a bubble.
That doesn't prove anything by itself. Bubbles are often discussed for years before they pop. But it tells you the fear is real.
4 ways AI is different from the dot-com bubble
Now the other side. Because if you only read the scary half, you'll make scary decisions.
No. 1: The leaders make real money, lots of it
In 2000, most of the hottest dot-com companies were unprofitable. Many, like Pets.com, never had a path to profit.
Today's AI spending is led by some of the most profitable companies in history. Microsoft reported its AI business passed a $37 billion yearly revenue run rate in early 2026, up 123% in a year. Alphabet's net income jumped 81% in the first quarter of 2026.
No. 2: Much of the spending is paid for out of profits
The dot-com build-out was largely funded by debt and by selling new shares to excited investors. Today's giants are mostly paying from their own cash flow, though borrowing is now rising fast, as you saw above. That's worth watching.
No. 3: The price tags are high, but not Cisco-high
Remember Cisco at more than 200 times earnings? Today's AI leaders are expensive, but most trade at a fraction of that. The CAPE for the whole market is near 2000 levels, but the most extreme dot-com prices were in a different league.
No. 4: There's real, paying demand
Google Cloud reported a backlog of more than $460 billion in early 2026. Microsoft reported commercial bookings of more than $600 billion still to be delivered. Customers have signed contracts. That's very different from "eyeballs" and website visits, which is how many dot-com companies were valued.
So which is it? Here's the honest verdict
The truthful answer is that both stories can be right at once.
AI can be a real revolution and AI stocks can still be overpriced. The internet was real in 2000. Cisco was real. Its investors still waited 25 years.
And there is another trap: being right too early is the same as being wrong.
On December 5, 1996, Federal Reserve Chairman Alan Greenspan wondered aloud whether "irrational exuberance" had pushed stock prices too high. The S&P 500 closed that day at 744.38.
Investors who sold on his warning then watched the index more than double, to 1,527 in March 2000. And here's the kicker: even at the very bottom of the crash in October 2002, the S&P 500 closed at 776.76. That was still higher than the day Greenspan sounded the alarm.
What this means for you You don't need to know whether AI is a bubble. You need a portfolio that survives either answer. That is exactly what the next two sections give you. |
What an AI crash could do to your index fund, in dollars
Let's make this real. Imagine a harsh but plausible "AI bust":
The 10 biggest stocks fall 50% (many dot-com leaders fell far more).
The other 490 companies fall 20% as fear spreads.
International stocks also fall 20%.
High-quality bonds hold their value.
Here's how a $500,000 portfolio would fare in that scenario, depending on how it's built. (These are illustrations, not predictions.)
How the $500,000 is invested | Drop | Loss in dollars | Gain needed to get back to even |
All in AI stocks or an AI theme fund | −50% | −$250,000 | +100% |
100% S&P 500 index fund | −33% | −$164,500 | +49% |
80% S&P 500 / 20% bonds | −26% | −$131,600 | +36% |
60% S&P 500 / 40% bonds | −20% | −$98,700 | +25% |
40% S&P 500 / 20% international / 40% bonds | −17% | −$85,800 | +21% |
Two things jump out.
First, a plain S&P 500 fund is far safer than betting on AI stocks directly. Roughly 57% of it sits outside the top 10.
Second, the biggest protection doesn't come from picking the "right" stocks. It comes from your mix. The difference between a −33% fall and a −17% fall is the difference between a scare and a retirement disaster.
Remember: a 50% loss needs a 100% gain just to break even. (Here's the math nobody shows you on recovering from a 50% loss.)
The 5 things index investors should do now
None of these steps requires you to guess the future. All of them work whether AI booms or busts.
No. 1: Find out how much AI you already own
Most index investors own far more AI than they realise.
If you have $100,000 in an S&P 500 fund and the top 10 companies are about 43% of the index, roughly $43,000 of your money sits in just 10 companies. Most of them are tied to the AI boom. Nvidia alone was about 7.6% of the index in a March 2026 filing, so about $7,600 of your $100,000.
Now add any extras: a Nasdaq-100 fund, a "tech" fund, an AI theme ETF, a few shares of Nvidia you bought on a tip. Every one of those holds the same giants again.
Do this tonight: list every fund and stock you own. Look up each fund's top 10 holdings on its fact sheet. Add up how much of your total money is in the same 10 names. If it's more than about 40%, you're making a bigger AI bet than an index investor needs to.
(Comparing funds? See S&P 500 vs Nasdaq 100 for the Long Run.)
No. 2: Rebalance back to your target
Booms quietly change your risk without asking you.
Say you set a 60% stocks / 40% bonds mix with $500,000 a few years ago. After a big run-up in stocks, it has drifted to $360,000 in stocks and $140,000 in bonds. That's now 72/28. You are taking more risk at exactly the moment prices are highest.
Rebalancing means moving it back. In this example, you'd shift $60,000 from stocks to bonds to return to $300,000 / $200,000.
Why this works Rebalancing forces you to sell a little of what has become expensive and buy what has become cheaper. It's "buy low, sell high" without any predictions. Many investors simply check once a year, or whenever their mix drifts more than 5 percentage points from target. |
If you can, rebalance inside tax-sheltered accounts or by directing new money into the part that's underweight, so you don't trigger a tax bill.
No. 3: Diversify beyond the top 10
Remember the "lost decade," when the S&P 500 returned about −1% a year from 2000 to 2009?
Here's what most people never hear. It wasn't a lost decade for everyone. Research Affiliates found that over the same 10 years, emerging-market stocks and real estate trusts (REITs) each returned about 10% a year, and US small companies about 3.5% a year. Investors who owned more than one corner of the market did fine.
You don't need to chase those exact winners. Next time it will be different ones. That's the point. You simply own more than one engine, so one stalling doesn't strand you.
Simple ways index investors do this:
A total world fund that holds thousands of companies across about 50 countries. (See VOO vs VT: Should You Own Just the S&P 500 or the Whole World?)
An international fund alongside your S&P 500 fund. Vanguard's 2026 outlook projects about 4–5% a year for US stocks over the next decade, versus about 4.9–6.9% for stocks outside the US. Forecasts are often wrong, but they show how much valuations matter.
High-quality bonds for the part of your money that must stay steady.
An equal-weight S&P 500 fund for part of your US money, if you want to reduce the top-10 tilt. It gives each company the same weight, so the giants count no more than the rest.
Keep it simple. Two to four low-cost funds is plenty. (See How Many Index Funds Should You Own? and The Lazy 3-Fund Portfolio.)
No. 4: Keep 2–3 years of spending money safe
This one matters most if you are retired or close to it.
The real danger in a crash isn't the fall itself. It's being forced to sell stocks at the bottom to pay your bills. Selling low locks the loss in forever.
So if you'll need money from your portfolio in the next few years, keep that money in cash, a money market fund or short-term high-quality bonds. A common rule of thumb is 2–3 years of planned withdrawals.
Example: you're 62 and take $40,000 a year from your investments. You'd keep about $80,000–$120,000 in safe, boring holdings. If AI stocks crash 50%, you spend from the safe bucket and leave your stocks alone to recover.
(Planning withdrawals? Read Will You Run Out of Money Before You Run Out of Life? The Simple "4% Rule".)
No. 5: Write a one-page crash plan, then put your investing on autopilot
The worst investment decisions are made at 2 a.m., in a panic, after a red headline.
So decide now, while you're calm. Write down what you will do when the market falls 20%, 30% and 50%. For most index investors, the answer is "nothing different, except rebalance."
Here is a simple decision path you can print and keep:
Then automate. Set up a fixed monthly investment into your chosen funds. If AI stocks crash, your same monthly amount buys more shares at lower prices. If they keep rising, you're still in. (More on this in Lump Sum or Monthly? How to Put Money Back Into the Market After a Big Loss and Who Else Wants to Invest Just 15 Minutes a Year.)
The 3 things that could ruin you
Now the warnings. Each of these has destroyed ordinary investors in every bubble in history.
No. 1: Going all-in on AI stocks, AI theme funds or leveraged ETFs
When something is soaring, it's tempting to move your money into "the thing that's working." That's how people ended up owning Cisco at $80 in 2000.
AI theme funds and single-stock bets concentrate you in exactly the part of the market most exposed to a bust.
Leveraged ETFs are even more dangerous. These funds try to deliver two or three times an index's return each day, and they reset daily. The SEC and FINRA both warn that over periods longer than one day, their returns can differ sharply from what buyers expect, and that they're generally unsuitable for long-term buy-and-hold investors.
Here's why, in two days of simple math:
Day 1 | Day 2 | After 2 days | |
The index | −10% | +11.1% | 0% (back to even) |
A 3x leveraged fund | −30% | +33.3% | −6.7% (still losing) |
The index went nowhere. The 3x fund lost money. Now picture that happening day after day through a volatile crash.
Morningstar found that investors in the most volatile funds lose far more to bad timing than investors in calm funds. Over the 10 years to 2025, the "behaviour gap" was about 2.1 percentage points a year for the most volatile funds, versus about 0.4 for the least volatile. Exciting funds make people buy high and sell low.
(If any of this sounds like you, read 7 Signs You're Gambling in the Stock Market, Not Investing.)
No. 2: Selling everything to "wait for the crash"
This feels safe. It's often the most expensive mistake of all.
The problem is that the market's best days tend to arrive in the middle of its worst stretches, when everyone is most afraid. J.P. Morgan found that from 2005 to 2024, seven of the S&P 500's 10 best days came within two weeks of one of its 10 worst days.
$10,000 left alone in the S&P 500 from 2005 to 2024 grew to about $71,750. Miss just the 10 best days and it became about $32,871. Less than half.
Even the professionals can't time it. Look at what happened to two of the most famous investors on Wall Street in early 2000:
Julian Robertson of Tiger Management refused to buy overpriced tech stocks. He was completely right. But his fund kept losing money and investors kept leaving, and he closed it on March 30, 2000, just weeks after the Nasdaq peaked. Right on the bubble, wrong on the timing.
Stanley Druckenmiller, who ran George Soros's Quantum Fund, knew tech stocks were overpriced. But, frustrated as they kept rising, he bought about $6 billion of them near the very top in March 2000. He later said he lost about $3 billion on that one move, and was out of Soros's firm by the end of April. Asked what he learned, he said: "I already knew that I wasn't supposed to do that."
If two legends with research teams couldn't time the top, a burned investor with a phone app shouldn't try. Morningstar estimates the average dollar in US funds earned about 1.2 percentage points a year less than the funds themselves over the 10 years to 2025, simply because of when people bought and sold.
(Two related reads: Why Smart, Hard-Working People Sell at the Very Bottom and If You've Ever Lost Money in the Stock Market, Read This: The Hidden Gap.)
No. 3: Borrowing to invest, or investing money you'll need soon
A fall in prices hurts. A fall in prices with debt attached can end your investing life.
When you buy on margin, your broker can demand more cash when prices fall. If you can't pay, they sell your shares for you, at the worst possible moment. That's why record margin debt (about $1.5 trillion in June 2026) worries so many analysts. Forced selling feeds on itself.
The same trap catches people without margin accounts. If the money for next year's tuition, a house deposit or your living costs is sitting in stocks, a crash can force you to sell at the bottom.
The rule is simple: never borrow to buy stocks, and keep money you need within about five years out of the stock market.
(If you've already been hurt this way, start with Can't Sleep After a Big Stock Loss? 9 Things to Do Before You Touch Your Portfolio Again.)
Your "do this tonight" checklist
Print this. Tick it off. It takes about an hour.
List every fund and share you own, with today's value.
Add up how much of your total sits in the same top-10 companies. Note the percentage.
Write down your target mix (for example, 60% stocks / 40% bonds). Compare it with your actual mix.
If you're more than 5 points off, plan your rebalance, ideally inside tax-sheltered accounts or with new money.
Check you hold more than one engine: international stocks, bonds, or a total world fund.
Count how much you'll withdraw in the next 2–3 years. Make sure that amount is in cash or short-term bonds.
Review any margin loans, and any leveraged or inverse ETFs you're holding as long-term investments. Regulators warn these aren't built for buy-and-hold.
Set up an automatic monthly investment.
Write your one-page crash plan and put it somewhere you'll see it.
Turn off the market news alerts on your phone.
Frequently asked questions
Is AI the next dot-com bubble?
Nobody knows yet. Today shares key warning signs with 2000: record market concentration, Shiller CAPE near 40 and record margin debt. But today's AI leaders are highly profitable, unlike most dot-com firms. AI can be a real revolution and its stocks can still be overpriced, as Cisco proved in 2000.
What happens to my S&P 500 index fund if the AI bubble bursts?
It would likely fall, because the 10 biggest companies, mostly AI-linked, make up about 43% of the index. In a hypothetical bust where they fall 50% and the rest fall 20%, an S&P 500 fund would drop about 33%. That's painful but far less than owning AI stocks alone.
Should I sell my index funds before the AI bubble pops?
For most long-term investors, no. Timing a top is extremely hard, and the market's best days often come right after its worst. Missing just the 10 best days from 2005 to 2024 cut a $10,000 S&P 500 investment's ending value by more than half. Rebalancing to your target mix is a calmer alternative.
How do I protect my index funds from an AI crash?
Check how much of your money sits in the top-10 companies, rebalance to your target mix, add international stocks and bonds, keep 2–3 years of withdrawals in cash or short-term bonds, and keep investing monthly. Avoid margin debt and leveraged ETFs.
Is an equal-weight S&P 500 fund safer in an AI bubble?
It's less concentrated, because every company gets the same weight, so the AI giants count no more than any other. That can soften a crash led by the biggest stocks. But it has its own risks, usually charges more, and will lag when the giants keep winning. Many investors use it for only part of their US money.
How long did it take to recover from the dot-com crash?
The Nasdaq Composite took about 15 years, regaining its March 2000 peak only in April 2015. The broader S&P 500 recovered faster but was hit again in 2008. Investors who owned a mix of international stocks, small companies, real estate and bonds did much better through the 2000s.
Should I buy AI stocks now that prices have dropped?
If you own a broad index fund, you already own the AI leaders in proportion to their size. Adding more increases your bet on one theme. If you choose to, keep it a small slice you could afford to see halve, and never use borrowed money.
The calm investor's edge
Bubbles are loud. They make headlines, they make some people rich for a while, and they make everyone else feel left behind or terrified.
But history has a quieter message. The investors who came through 2000 best weren't the ones who guessed the top. They were the ones who owned a sensible mix, kept money aside for their needs, refused to borrow, and kept buying, month after month, while everyone else panicked.
You can do all of that, starting tonight. Whether AI turns out to be the next internet or the next Pets.com, a prepared index investor wins either way.
If you've lost money in the market before and want to understand exactly why it happened, so it never happens again, get my free 38-page report, "Why You Lost Money in the Stock Market (And Why It Wasn't Your Fault)." It's plain English, free and an instant download. Inside, you'll discover:
The seven Wall Street traps that quietly drain ordinary investors.
How a "small" yearly fee can eat a huge slice of a nest egg over time.
Why most professional fund managers trailed a simple index over 20 years.
The "boring" strategy Wall Street hopes you never discover.
Whether it's too late for investors over 40 (it isn't).
About the author
Stephen Loke has been a student of the market for more than 15 years. He is also an agritourism entrepreneur whose farm has been featured in Bloomberg News and Asahi Shimbun.
This article is general education, not personal financial advice. I'm not a licensed financial adviser. Figures are from the sources below and were checked in November 2026; markets change daily. The crash scenario is a hypothetical illustration, not a forecast. Think about your own situation, or speak with a licensed professional, before making investment decisions.
Sources
AP via Fox19: A brutal sell-off for AI stocks sends Wall Street reeling (Nov. 5, 2026)
RBC Wealth Management: The "Great Narrowing": S&P 500 concentration (Jan. 2026)
InvestmentNews: Is the S&P 500 still diversified? (July 2026)
ChartRow: S&P 500 concentration, from fund N-PORT filings (Mar. 2026)
Silicon Canals / Financial Times tally: Big Tech capex $410B in 2025, ~$725B planned for 2026
OfficeChai: Hyperscaler capex and Bank of America debt forecast
Yahoo Finance: Hyperscalers hit $700 billion in 2026 AI spending plans
Resultsense / FT: Alphabet Q1 2026 results and cloud backlog
Advisor Perspectives: FINRA margin debt record $1.53 trillion (June 2026)
Insurance Journal / Bloomberg: "AI bubble" news mentions (Jan. 2026)
The Irish Times: Bank of America fund manager survey on AI bubble (Oct. 2025)
Bloomberg: Cisco shares finally top dot-com record after more than 25 years (Dec. 2025)
Phys.org / AP: After 15 years, Nasdaq recoups losses of dot-com crash (Apr. 2015)
Benzinga: Greenspan's "irrational exuberance" speech, S&P 500 close 744.38
Institutional Investor: Julian Robertson closes Tiger Management (Mar. 30, 2000)
Slashdot / Financial Times: Cisco's P/E above 200 at its 2000 peak
Bloomberg via Auburn University archive: Druckenmiller and Roditi resign (Apr. 28, 2000)
The Acquirer's Multiple: Stanley Druckenmiller on his biggest mistake
Voya: J.P. Morgan analysis of missing the best days, 2005–2024
401(k) Specialist: J.P. Morgan Guide to Retirement on best and worst days
Morningstar: Mind the Gap, investor returns to Dec. 31, 2025
Advisor Perspectives: Volatility and the investor return gap (Aug. 2026)
SEC Investor.gov: Updated Investor Bulletin, Leveraged and Inverse ETFs
FINRA: Leveraged and inverse ETFs, extra risks for buy-and-hold investors
Yahoo Finance / Motley Fool: Vanguard 2026 outlook, US vs international expected returns
Comments