If you have a 401k, you are one of the biggest investors in the AI boom. As AI stocks climbed, so did your retirement balance, whether you picked those stocks or not.
Now the people who built the boom are questioning it. Mark Zuckerberg, whose company Meta has poured billions into artificial intelligence, was asked about the bubble talk and called it "quite possible."
The White House went further. An internal report called the AI boom dangerous and warned that if it pops, the damage could reach far past the economy and into millions of Americans' 401ks.
That warning is a reason to look inside your account, not a reason to panic. Knowing where the money sits is how you protect it, and it's also how you spot the opportunities that show up when prices swing.
Part of that opportunity sits in the dollar itself. Our CEO Jaspreet Singh is hosting a free live investor workshop on September 29th on how to profit when the dollar is losing value, and you can save a free seat here.
Michael Burry Says the AI Bubble Looks Like 1999
Big-name investor Michael Burry posted on X that this feels like the last months of the 1999 to 2000 dot-com bubble. Back then, internet stocks kept climbing because everyone wanted a piece of the next big thing.
The bubble burst in 2000, and internet stocks fell 78%. The worry now is that AI does the same thing, and this time takes 401ks down with it.
$10 Trillion of 401k Money Is Feeding the AI Bubble
The US stock market is worth about $77 trillion. Over the same decades that number exploded, 401k accounts grew right alongside it, and today they push roughly $10 trillion into stocks and other investments.
Two places soak up most of that money:
- Target date funds, the number one 401k investment, which are built to grow your money toward a set retirement year.
- S&P 500 funds, which hold the 500 largest companies in the US stock market and rank among the biggest investments outside 401ks too.
Dig into a target date fund and you find it is mostly a basket of other funds. One of the biggest baskets is the S&P 500, so both roads lead to the same 500 stocks.
One of Every $3 in Your 401k Lands in Five AI Stocks
"You own all 500 companies" is true, but it hides the weighting. Put a dollar into an S&P 500 fund today and the first stops for that money look like this:
| Company | Share of each S&P 500 dollar |
|---|---|
| Nvidia | 7.9% |
| Apple | 7.0% |
| Alphabet | 5.5% |
| Microsoft | 5.4% |
| Amazon | 3.9% |
| Total | About 30% |
That is one out of every $3 you invest going straight into five stocks. All five are tech companies, and all five are deep in AI.
The arrangement works beautifully while AI is booming. It also means most 401k holders carry a heavy bet on a single industry without knowing it, and that kind of concentration can hurt when the industry turns.
The usual answer is "my target date fund is different, I'm diversified." One of the most popular is the Vanguard Target Retirement 2055 fund, and its top holdings are Apple, Nvidia, Microsoft, and Amazon.
This setup has made a lot of investors rich on the way up. The question is what it does on the way down.
Four Ways the AI Bubble Is Already Bigger Than the Dot-Com Bubble
History doesn't repeat itself, but it rhymes. Four measurements from the run-up to 2000 are bigger today.
| Signal | Before the 2000 crash | Today |
|---|---|---|
| Top 10 companies' share of the S&P 500 | 27% | About 40% |
| Tech's share of the S&P 500 | About a third | 38% |
| Buffett Indicator | About 140% | About 240% |
| Index funds' share of the stock market | About 6% | About 54% |
Ten companies now make up 40% of the index
Before the dot-com bust, internet companies took all the gains while everyone else was left out, and the top ten names made up a quarter of the S&P 500. Today those ten, mostly tech, are closer to four-tenths of the whole index.
Tech is 38% of the S&P 500
A third of the index was tech heading into 2000, and today it is 38%. Part of that is fair, since technology itself has grown and the economy needs more tech companies, but it is a number worth watching.
The Buffett Indicator is at 240%
The Buffett Indicator measures how expensive the stock market is compared to the real value of the businesses underneath it. It hit around 140% before 2000, when stocks were already pricey, and it sits around 240% today.
In plain terms, investors are paying far more for every dollar of profit than they did last time.
Index funds own 54% of the market
Index funds, which buy every stock on a list instead of picking favorites, were a small slice of the market in 2000, at about 6%. Today they are 54%.
When index funds were small, a person chose each stock, and that person might skip Microsoft, want Alphabet, or decide they already owned too much tech.
Now most investors buy the S&P 500 or a target date fund because it has performed amazingly, and they hope it keeps going. Enormous sums flow into stocks that investors don't even know they own.
Why the AI Bubble Isn't a Repeat of 2000
In the late 1990s, many internet companies had no profit and no revenue at all, and they got huge valuations on nothing but a business model. Today's AI leaders make real revenue and real profits.
The other difference is what happened after the crash. The dot-com bust didn't kill the internet, and while many companies went bankrupt, internet companies grew again and remain some of the largest in the world.
Even if the AI bubble bursts, some companies will go away, but AI itself won't.
Circular Financing Is the Risk 2000 Never Had
One difference cuts the other way. Circular financing, where companies fund each other in a loop, wasn't really part of the 2000 story.
Nvidia, the biggest company in the world, invests money in a company or buys its products. That company then turns around and buys chips from Nvidia.
Multiply that across a group of tech companies and you get a set of firms fueling money into one another. If one domino stops, the whole house of cards could go with it.
The White House Is Pumping Air Into the AI Bubble
Is having your retirement tied to AI a bad thing? For years AI has been the biggest driver of the economy and the biggest gainer in the stock market, and that gain is what lifted 401ks.
Now it is being supercharged: the White House does not want to lose the AI race to China, so the government is pouring billions into artificial intelligence. It has also issued executive order after executive order to make it easier for AI companies to expand in the US.
Private money plus government money means tech and AI companies boom, and your retirement grows with them. The concerns are oversaturation, meaning too much money chasing AI, and the amount of debt funding all that investment.
A boom and a bubble can happen at the same time. The dot-com era was both: a real economic boom that grew too big, too fast, then right-sized so hard that internet stocks fell 78% before booming again.
AI is changing the economy the way the internet did, and it will keep changing it.
The AI Bubble Is Also an Arms Race With China
The US is fighting to stay the world's superpower and to keep the dollar strong. That fight is why it can't afford to lose the AI race, and it's the same reason so many investors worry about the dollar losing value.
Turning that worry into a plan is the whole point of the free live investor workshop our CEO Jaspreet Singh is hosting on September 29th. It covers how to profit when the dollar is losing value instead of getting squeezed by it, and you can register free here.
| Metric | United States | China |
|---|---|---|
| Forward P/E ratio | 22x | About 13x |
| AI investment in 2025 | About $286 billion | About $12 billion |
| Best AI model, head to head (2025) | Slightly ahead | Slightly behind |
| Cost to run 1 million tokens on the best model | About $15 | About $0.55 |
The forward P/E ratio is the price investors pay for every dollar of profit a company is expected to make. US stocks trade at 22 times earnings versus about 13 times in China, so American profits cost roughly 70% more.
China's numbers aren't always reliable, because its government doesn't always share them. On the best information available, though, the US outspent China on AI by roughly 24 to 1 in 2025.
The US model scores slightly higher head to head, but running it costs about $15 per million tokens, the units of text an AI model processes. China runs its best model for about $0.55 per million, so America's AI is slightly better and a whole lot pricier.
Winning that race comes down to resources, and the resource is money, from private investors or from the government. Nobody can say what Washington will do, but it does not want to lose, and it could pour in even more.
Should You Sell Your 401k Before the AI Bubble Pops?
Nobody online can tell you what to do with your retirement account. Investing has risks, you will lose money at some point, and the due diligence has to be your own.
Still, a lot of investors hear all of this and want to dump their AI stocks and sell out of their 401k. Before doing that, consider that the last 100 years brought 16 recessions and 25 market crashes, so the next one is a matter of when, not if.
You wanted broad exposure to the market and got a fund heavy in AI, because AI is heavy in the economy right now.
If you are a long-term investor, that is the game you signed up for. When the AI bubble deflates you will feel the pinch, but you only lock in the loss if you sell, and the broad economy grows again even if it takes years.
Investing through the crashes is the whole point. An investor who can't do that shouldn't put money into the stock market at all.
Age changes the math. Investors in or near retirement are often told to pull some money out of stocks, because a crash could hit before they have time to earn it back.
Younger investors get the opposite advice, because downturns create the biggest buying opportunities of all time:
- 2022: stocks fell 20%
- 2020: stocks fell 34% when the pandemic hit
- 2008: the market fell 50%
- 2000: internet stocks fell 78%
Every one of them was a great time to buy. Buying individual companies in a downturn takes more work, more research, and more risk, because any single company can go bankrupt.
A broad fund like the S&P 500 isn't going away, the same way the internet didn't. That is where ABB comes in: always be buying, and buy even more aggressively when the market drops, because the whole economy is on sale for 50 cents on the dollar.
With ten or 15 years ahead of you, the bubble is almost beside the point. Enjoy it on the way up and buy the window on the way down.
The investors who get hurt are the ones chasing the hype up and the panic down. They sell their retirement funds in a downturn and turn a long-term investment into a day trade.
Emotional investing is the enemy of profits. Know your strategy and follow it to the end.
When AI stocks do fall, the news will point at the Nvidia feedback loop, the government money, and the 401k money that inflated these companies. The investors who understand that now won't be learning it for the first time in a panic.
