AI boom or bubble? How to navigate risks and opportunities

Investors would be forgiven for feeling confused. They are surrounded by gloomy predictions that artificial intelligence (AI) is a bubble, with unsustainable spending and lofty valuations. Yet parts of the technology sector keep powering higher, with NVIDIA Corp (NASDAQ:NVDA) recently reporting an 85% increase in revenues year on year. The technology story is evolving and investors may need to shift their approach to keep pace.

Hedge fund manager Michael Burry, who successfully called the 2008 financial crisis, said in a Substack post: “History is not a perfect guide, but I see so many indicators both technical and fundamental lining up for the same conclusion…It’s an asset bubble, pure and simple.”

Burry is far from alone in adopting a bearish stance. A recent study from Janus Henderson found that more than two-thirds of investors are concerned about an AI bubble or market correction over the next year.

• Invest with ii: SIPP Account | Stocks & Shares ISA | See all Investment Accounts

Yet the same survey found that 61% expected AI to have a positive long-term impact on markets. There are plenty who believe that this is a genuine revolution, likely to be more important and disruptive than the internet. Their view is that, if anything, markets are underestimating its impact. Morningstar, for example, believes that in spite of Nvidia’s huge share price rise over the past five years (up over 1,000%), the AI sector could still be as much as 20% undervalued.

Mark Preskett, senior portfolio manager at Morningstar, says the trend has been evident in the latest round of quarterly earnings announcements: “Companies are cutting staff and investing in AI. It’s transcending sectors. It’s not just software firms, it’s accountancy firms, it’s banks, it’s insurance companies, it’s data providers. It’s the whole business world. We’re seeing incredible earnings growth from a cohort of stocks.”

Which side to believe? The answer is more nuanced than the “bubble or no bubble” divide. Gavin Haynes, a consultant at Fairview Investing, says: “There is no doubt that AI is transformative and provides exciting growth opportunities. However, as in any technology revolution there will be winners or losers and the question is whether the current euphoria driving markets higher is justified.”

• AI boom is not a bubble, and why we don’t own defence stocks
• How to spot (and survive) a market bubble 

James Anderson, who for a long period was the fund manager of Baillie Gifford’s Scottish Mortgage Ord (LSE:SMT) investment trust, told investors in his annual letter that the top US internet and software stocks were vulnerable as AI spending cut their cash flows.

Rather than the “hyperscalers” - Google (Alphabet Inc Class A (NASDAQ:GOOGL)), Meta Platforms Inc Class A (NASDAQ:META), Amazon.com Inc (NASDAQ:AMZN) and Microsoft Corp (NASDAQ:MSFT) - that have reaped the spoils of cloud computing and the internet, it will be the hardware giants such as Nvidia, Taiwan Semiconductor Manufacturing Co Ltd ADR (NYSE:TSM) and ASML Holding NV (EURONEXT:ASML) that win in this cycle. 

Premium valuations, but are they justified?

Clare Pleydell-Bouverie, manager on the Liontrust Global Innovation C Acc GBP (B8DLY47) fund, agrees that while this is not a bubble equivalent to the dotcom boom and bust of the late 1990s, the winners of the previous cycle will not be the winners of this cycle. 

“We get asked a lot whether we are in an AI bubble. When you see infrastructure spending of the type we’re seeing now - $700 billion (£523 billion) in capex from the hyperscalers this year and likely to go up next year - there are plenty who believe we’re in a cycle of over-building. The fundamental difference is that in 1999, the capital went in first and the revenue was supposed to follow: the demand didn’t come in time. Today, every unit of capacity installed, there is demand for it and it’s been monetised instantly.”

The biggest question is around the fortunes of the hyperscalers, which have seen their buoyant cash flows disappear on AI spending. Pleydell-Bouverie believes they should be able to reap the rewards from their investment in the longer term, but they may not be as important in future. Equally, some of the software companies are seeing a genuine challenge to their business. The S&P Software & Services Select Industry Index, a proxy for the software sector, dropped more than 25% between mid-January and mid-February.

• Sign up to our free newsletter for investment ideas, latest news and award-winning analysis
• The most-popular ISA funds today vs five years ago 

This nuance matters when valuations are so high. Haynes says: “Although it is very different to the dotcom bubble, valuations are looking demanding - the Nasdaq trades over 40 times - and it feels prudent to be selective in where you invest and be aware that US, Global and Asian passive funds will already be giving you significant exposure to the sector.” 

The tech sector now makes up 27% of the MSCI World Index, 33% of the S&P 500, and 41% of the MSCI Asia Ex Japan Index, so most investors with global equity exposure are likely to have a material weighting.

Sheridan Admans, founder and chief investment strategist at Infundly, points out that most people will have plenty of exposure anyway: “The AI question is no longer simply whether to be bullish or bearish. Most portfolios already have meaningful exposure through global equity funds, US equity funds, technology trusts, technology thematic exchange-traded funds (ETFs) and quality-growth strategies. I’m finding the more useful question is whether that exposure is deliberate, diversified and sensibly valued.”

Fund routes

A technology index has historically been a reasonable choice to get exposure. Haynes: “A passive fund tracking the sector has been hard to beat as the largest companies have been offering some of the best returns.” However, as AI disrupts and new winners emerge, the index will include companies on the wrong side of the disruption as well as the right side. There may be better options to take more targeted exposure.

Investors are likely to have a lot of exposure to the hyperscalers already, and with their dominance less certain, probably don’t need more. Haynes adds: “AI is changing so quickly that managers need to be flexible and not just buy and hold.”

For an active option, Haynes says Polar Capital Technology Ord (LSE:PCT) is hard to beat. The trust is up 2,224% over the past three years and – inexplicably – remains at an 8.8% discount to net asset value (NAV), according to Morningstar on 5 June 2026.

Its peer, Allianz Technology Trust Ord (LSE:ATT) has gains of 190.4% over three years. Its discount is 7.4%. Historically, it has ventured more into mid-caps versus PCT.  For those wondering why there’s a sizeable performance gap between the two, a recent article by interactive investor examined why PCT has pulled ahead of ATT.

Admans likes BlueBox Global Technology R UK GBP Acc (BSS9B56) for a focused technology allocation. “It is interesting because it is not simply chasing the most obvious mega-cap AI winners. Its philosophy is built around ‘direct connection’ companies enabling computers to sense, interpret and act on the real world through semiconductors, sensors, automation, robotics and connected devices. That gives it exposure to the infrastructure behind AI and digitalisation, while its persistent mega-cap underweight helps distinguish it from many benchmark-led technology options.”

• Are higher rates a red herring for the tech rally?
• Fund Focus: an ETF portfolio to beat the market?

Other options include the Matthews China Innovators I GBP Acc (BJN4L97) fund. This has over a third of its holdings in technology, but also a range of other growth themes. Chinese technology companies are generally less highly valued than their Western peers. Preskett also believes Asia is a good hunting ground: “Our view is that South Korea remains the neatest way to access this trade. Two of the three main memory manufacturers are there.” Barings Korea I GBP Acc (B9M3RQ4) is an active option, or there are a range of passive choices, such as the iShares MSCI Korea ETF USD Dist GBP (LSE:IKOR).

For those who remain unconvinced, and believe a bubble remains a possibility, there are some natural counterweights to the AI trend. Admans says: “A fund like the Redwheel Global Intrinsic Val R GBP Acc (BJBPX96) is not ‘anti-AI’  in a literal sense, but it is a useful yin to the technology yang. The strategy is highly active, valuation-led and materially underweight US mega-cap growth, with greater exposure to cheaper, out-of-favour regions, mid-caps and cyclical recovery opportunities.”

For Admans, an active technology fund and a more value-focused fund frame the current debate well. He adds: “Investors can participate in the structural AI opportunity, but balance it with areas where expectations and valuations are less dependent on the AI narrative continuing perfectly.”

Haynes agrees that moving into other areas such as value and equity income funds can provide genuine diversification. “Also, from a regional perspective, Europe and the UK have relatively low tech weightings”. UK equity income funds, for example, will have almost no technology exposure.

The “bubble” debate may be misleading. Some of the AI infrastructure companies have seen a strong run, but could still be undervalued given the vast spending going into the build-out of AI capabilities. However, there are questions round the hyperscalers and the software companies. This argues for a more nuanced approach to technology investment from here.