The biggest investment risk surrounding artificial intelligence may not be whether the technology succeeds, but whether investors have unknowingly become overexposed to it.

That is the warning from InvestmentMarkets CEO Darren Connolly, who says the rise and broad impact of the thematic and the sheer volume of investment options has quietly reshaped millions of investment portfolios, often without investors realising just how concentrated their holdings may have become.

"Most people think they have a diversified portfolio because they own global funds or ETFs," Connolly said.

"But if you look under the bonnet, and particularly if you have multiple investments, you may be holding the same handful of companies multiple times. Nvidia, Microsoft, Amazon, Alphabet and Meta dominate global equity indices, which means AI exposure has become embedded across countless portfolios."

The extraordinary growth of AI infrastructure businesses and the number of investment products exposed to the theme has reignited debate over whether markets are entering another technology bubble.

Unlike the dot com era, many of today's market leaders are highly profitable businesses generating substantial cash flows. Earnings have continued to rise alongside share prices as demand for semiconductors, cloud infrastructure and AI computing accelerates.

That makes today's environment far more complicated than previous market manias.

"These businesses are delivering real earnings growth and investing heavily because demand is real. That is what makes it so difficult to determine whether we are looking at a bubble or the early stages of a genuine technological transformation."

The greater concern, he says, is what happens when a thematic and small group of companies begins driving an outsized share of market performance.

Technology and communication services companies now account for around half of the S&P 500, while the ten largest companies make up close to 40 per cent of the index.

"Investors may believe they are diversified because they own hundreds of companies through broad market funds. In reality, portfolio performance may be becoming increasingly dependent on a single thematic."

Countries including Taiwan, South Korea, and Japan have become central to the AI supply chain through semiconductor manufacturing, advanced memory technology, robotics and computing infrastructure.

As a result, regional diversification may no longer reduce AI exposure as effectively as investors expect.

"The AI story has become global. You can own United States equities, Asian equities and technology funds and still be relying on exactly the same investment theme."

One of the biggest questions now facing markets is whether the enormous capital being committed to AI infrastructure will ultimately generate returns that justify the investment.

Technology giants are expected to spend hundreds of billions of dollars building AI capability over the coming years, with some forecasts projecting AI related infrastructure investment reaching trillions of dollars before the end of the decade.

Rather than attempting to predict when enthusiasm for AI will peak, Connolly believes investors should focus on understanding how much exposure they already have and whether their portfolios remain appropriately balanced.

"History tells us that transformative technologies create extraordinary opportunities, but they can also produce periods of excessive optimism," he said.

"The question isn't really whether AI will change the world. The more important question for investors is whether they are paying the right price for the opportunity and whether they have enough diversification in their portfolio if market sentiment turns."