AI stocks are racing ahead ‘almost in isolation’ — and the bubble could burst. Experts say investors should be worried

A new concerned voice has joined the likes of JPMorgan Chase CEO Jamie Dimon, Mad Money host Jim Cramer and The Big Short inspiration Michael Burry in sounding the alarm on this worrying market trend. Jim Paulsen recently flagged a troubling trend in the S&P 500 that anyone with money in North American markets should know about.

‘Extreme’ division between new and old era stocks

Paulsen spent decades as chief investment strategist for the Leuthold Group and now shares his views through a Substack newsletter followed by thousands of readers. His recent posts focus on what he calls an “extreme” split of the market, one that he says isn’t good news for AI investors.

As Paulsen explains, what has historically kept stock market rallies on solid ground is the participation of “old era” stocks — banking, manufacturing, energy and consumer goods — which tend to trend in the same direction as the tech stocks driving the gains. When they go their separate ways, things tend to get ugly.

But what we’re seeing now is the opposite: AI shares that are “racing ahead almost in isolation,” which Paulsen suggests is an almost guaranteed sign of trouble.

“For the last 30 years, the correlation of daily price movements between new era and old era stocks during the last year has proved to be a good risk indicator for new era investors,” Paulsen wrote in his recent post.

“The most recent rally in new era stocks since March 30 has been explosive, causing a breakout of optimism among investors that AI excitement is leading the stock market on another significant leg higher. However, this latest rally has been associated with an alarming drop in the trailing 12-month new/old era stock price correlation, suggesting the contemporary rally may not be sustainable.”

As Paulsen’s research shows, when this pattern has appeared in the past, the stocks that had been driving market gains typically slowed or dropped — what he calls a “notable pause,” if not “meaningful underperformance.”

Get a brokerage account that fits your needs. You can't control the market, but you can control fees, tools and how you invest. Find an online investing platform that helps you invest with confidence. Compare Canada's Best Brokerages.

The market split has reached a drastic level

Bifurcation (when a market splits into two groups moving in opposite directions) isn’t new. But the gap between AI-driven stocks and everything else has grown remarkably wide.

The divide began in 2022, when the current bull market took root. But it has become much more pronounced over time. In the most recent 28-session rally between late March and early May 2026, just 10 stocks drove 69% of the S&P 500’s total gains, according to data from Investing.com.

Even more striking: New era AI stocks have performed, on average, nearly seven times better than the rest of the S&P 500 index since year-end, up 36.2% versus 5.3% for old era stocks. In mid-May, 5% of S&P 500 components sank to 52-week lows while the overall index was at a record high — only the fourth time in recorded history this has ever happened.

Meanwhile, the 10 largest companies in the S&P 500 now account for roughly 40% of the entire index by market cap. This level has not only matched, but exceeded the peak concentration seen during the dot-com bubble of 2000, when the top 10 held around 27%.

As Paulsen himself puts it, even if the reasoning for why some AI companies are outperforming is valid, the question instead becomes “how sustainable a bull market is where most companies are essentially failing.”

Paulsen’s warning is echoed at the institutional level. A Deutsche Bank fund manager survey found that 57% of institutional investors now identify an AI valuation crash as the single greatest risk to markets.

What this means for Canadian investors

It would be easy to dismiss this as an American problem. After all, the Toronto Stock Exchange (TSX) is structured quite differently from the S&P 500, with heavier weightings in energy, financials and materials.

In fact, for only the third time in the past 15 years, Canada’s S&P/TSX Composite Index generated a higher annual return than the S&P 500 in 2025, posting a gain of over 28% compared to approximately 18% south of the border.

But Canadian investors are far from insulated.

Canadian-listed AI-related stocks have surged dramatically. Celestica Inc., a Toronto-based data centre infrastructure company, climbed from under $80 a share to nearly $390 a share during 2025, and is currently trading around $500 as of July 7, 2026.

More broadly, millions of Canadian investors hold S&P 500 index ETFs inside their TFSAs and RRSPs — but the old “just buy the index” advice is being questioned, and for good reason. When the index is dominated by only a handful of big tech companies, a drop in a few of those stocks can drag the whole index down with them, and investors who simply track the index have nowhere to hide.

For Canadians holding U.S. stocks in a TFSA or RRSP, the exchange rate adds another risk to consider. If U.S. markets drop sharply, any losses could look even worse when converted back to Canadian dollars, depending on where the exchange rate sits at the time.

Canadian-listed AI ETFs — such as the CI Global Artificial Intelligence Fund ETF, Canada’s largest dedicated AI equity fund with nearly $1 billion in assets, or the TD Global Technology Leaders Index ETF, with a $3.5 billion portfolio — offer exposure to the same AI mega-caps Paulsen and Burry are warning about.

Your own portfolio comes down to risk tolerance

Joining the AI party can indeed produce big wins. The TSX’s Celestica delivered triple-digit returns for three consecutive years through 2025.

But experts like Burry have warned that the sector is building far too much infrastructure, driven more by hype than by actual paying customers.

Goldman Sachs estimates that US$7.6 trillion will be spent on AI infrastructure between 2026 and 2031, with big tech companies alone on track to spend close to $527 billion to $765 billion on data centres and AI in 2026. But as Paulsen puts it, the real question is whether the stocks getting the most attention have already priced in a decade of that growth.

As Burry wrote on his earlier Substack in May, “stocks are not up or down because of jobs or consumer sentiment. They are going straight up because they have been going straight up [based] on a two-letter thesis (AI) that everyone thinks they understand.”

Whether you believe AI is the start of a new industrial revolution or simply a trend that will eventually calm down, its impact on markets has been massive and unlike anything we’ve seen before. But how much you believe in AI’s future — and how much of your money you want to bet on it — is ultimately a personal decision, shaped by how long you plan to invest, how much risk you’re comfortable with and your goals.