26 Feb Should I start to worry about an AI bubble?

Q. My portfolio is mostly in index funds, allocated over different sizes for diversification. But I’m starting to worry about this AI bubble people keep talking about. Should I pare back on the funds that hold these tech stocks, so should I just let it ride and stop looking?
— Investor

A. It’s certainly a timely question.

Artificial Intelligence (AI) has clearly driven pockets of extreme valuation and speculative behavior.

But whether it is a full-fledged “bubble” or just a frothy phase layered on top of real earnings growth depends on which part of the market you look at, said Gerard Papetti, a certified financial planner and certified public accountant with U.S. Financial Services in Fairfield.

A very small group of mega-cap AI leaders — Nvidia, Microsoft, Alphabet, Meta, Amazon, Broadcom and Tesla — have contributed a disproportionate share of index returns since late 2022, with AI-linked names responsible for the majority of S&P 500 gains and earnings growth, Papetti said.

In 2025, the S&P 500 returned 17.8%, and estimates suggest that Nvidia alone accounted for more than 15% of that gain, with the broader “Magnificent Seven” responsible for more than 40% of the total return, he said.

Outside these leaders, the S&P 493 has delivered much more modest returns and slower earnings growth, which implies that “AI euphoria” is concentrated rather than market-wide, he said.

Papetti said AI-related firms have accounted for roughly 75 to 80% of S&P 500 returns and earnings growth since ChatGPT’s launch, a level of concentration reminiscent of late-stage sector booms.

Capital spending is extraordinary: the big cloud “hyperscalers” are on track to spend on the order of hundreds of billions of dollars annually on AI data centers and chips, with estimates that AI-driven capex alone could reach around 1.5% to 1.6% of U.S. GDP in 2026, he said.

And venture capital (VC) flows are heavily skewed toward AI and machine-learning (ML) startups, with some reports suggesting nearly two-thirds of U.S. VC deals going to AI/ML in 2025, up from roughly one-quarter just two years earlier, he said.
Based on all of that, the AI sector may be entering a speculative phase, capital overbuilding risk and extreme index concentration, Papetti said.

But at the same time, several factors differentiate today from prior pure bubbles such as the late 1990s dot-com boom.

“The leading AI platforms are already highly profitable, with strong free-cash-flow support and tangible demand for AI chips and cloud services; their earnings growth so far has, at least in part, kept up with valuation expansion,” Papetti said.

The current environment may be a mix of “extreme bubble” conditions in certain names, where revenues are not driving profits and/or free cash flow and an opportunity for those companies that are driven by real productivity gains and durable business models, he said.

Rather than a market-wide bubble, there is a legitimate, earnings-backed cycle in the leaders, alongside speculative excess in smaller, less-proven AI plays, he said.

“AI has created a valuation bubble in parts of the equity market and may be a disruptor to certain industries as evidenced by the recent correction in Software as a Service (SaaS) sector,” he said. “AI has heavily distorted index leadership, but it has not created a uniform bubble across all stocks.”
“From a portfolio allocation standpoint, the main risk is concentration in a narrow group of AI winners whose pricing includes very optimistic long-term assumptions and most likely `priced-for-perfection,’” he said.

Email your questions to .

This story was originally published in February 2026. 

NJMoneyHelp.com presents certain general financial planning principles and advice, but should never be viewed as a substitute for obtaining advice from a personal professional advisor who understands your unique individual circumstances.