The tech optimism that has driven London’s FTSE 100 to record highs is showing cracks. A chorus of City analysts now warns that the frothy valuations of artificial intelligence stocks have created a bubble reminiscent of the dot-com era. With British investment funds heavily exposed to US-listed AI giants like Nvidia and Palantir, the warning is a cold splash of water for portfolio managers across the Square Mile.
The alarm was sounded by HSBC’s global equity strategists who note that AI-related equities now trade at a premium of over 40% compared to the broader market. The MSCI World AI index, a benchmark for the sector, has surged 180% since 2023, far outpacing earnings growth. This disconnect, analysts argue, is unsustainable.
“We are not saying AI is a fad. It is transformative. But the market has priced in a decade of perfection in six months,” says Leila Akhtar, technology analyst at Barclays. She points to the ‘Jevons paradox’ in AI: as models become more efficient, the demand for compute may drop, threatening the revenue models of hardware suppliers.
Adding to the unease, the Bank of England’s Financial Policy Committee has flagged the concentration of AI investments in passive funds. Over 60% of new UK retail money flowing into tech ETFs since 2024 has gone into funds with over 20% allocation to AI stocks. “If sentiment turns, the exit door will be narrow,” warned a committee member.
The bubble psychology is evident in the London markets. Shares in Darktrace, the UK’s cybersecurity AI darling, have tripled since January despite revenue growth of only 22%. Brokers now rate the stock as ‘hold’ at best. Similarly, Ocado’s AI-driven warehouse logistics division has delivered a 50% premium to its grocery business. Analysts at Bernstein call this “a stretched narrative.”
For British investors, the risk is acute. The UK’s pension funds, which hold over GB pound 2 trillion in assets, have increased their AI exposure fivefold since 2022, according to the Pensions and Lifetime Savings Association. With interest rates still restrictive and inflation sticky, a sharp correction could trigger margin calls and forced selling.
What could burst the bubble? Regulatory action is the most likely catalyst. The EU’s AI Act, effective April 2025, imposes compliance costs that could squeeze small AI firms. In the US, the SEC is probing whether AI stocks have been hyped by social media influencers and ‘pump and dump’ schemes. A jolt from either side could puncture euphoria.
But not everyone is panicking. The AI infrastructure story has legs: Microsoft, Meta, and Alphabet have committed over GB pound 100 billion in capital expenditure for 2025 alone. Data centre builders like Schneider Electric and Rolls-Royce’s small modular reactor division are genuine beneficiaries. The trick, say contrarians, is to separate the wheat from the chaff.
“The second-order effects of AI are real. It is the third-tier companies that will suffer,” explains Dr. Anil Patel of the Alan Turing Institute. He advises UK investors to look at firms deploying AI to solve specific, quantifiable problems: logistics optimisation, drug discovery, climate modelling. “Avoid the picks and shovels euphoria. Focus on the miners, not the equipment suppliers.”
The FTSE 100 has already shown jitters, falling 2.3% this week on AI profit-taking. The VIX, Wall Street’s fear index, is up 30% in March. The message is clear: for British investors who have ridden the AI wave, now is the time to check seat belts.
As one City trader put it, “The AI bubble will not pop with a bang, but a long, slow hiss of realisation that not every company with ‘deep learning’ in its pitch deck is the next Google.” The lesson from history is that bubbles burst faster than they inflate. British pensioners and retail investors alike should prepare for turbulence.
