A revolutionary new technology promises to change forever the way humans live and work.

Financial markets respond by soaring to record-breaking highs, making investors a fortune. Then it all comes crashing down. This is not the near future; it’s the year 2000.

This technology – the internet – would go on to become the foundation of the world’s economy, but that did little to protect those with money invested in the stock market from huge losses thanks to a handful of wildly over-valued internet start-ups.

There is now a growing fear that the bullishness over artificial intelligence (AI) is playing out in an eerily similar fashion to how the world reacted to the birth of the internet decades ago – and that a similar crash could bring even more devastating consequences.

‘Enormous implications’ for retirees

The bursting of the dot-com bubble resulted in the Nasdaq stock index, where American tech firms are listed, crashing by almost 80pc from its peak in March 2000 to its nadir in October 2002.

Anyone who was invested in the S&P 500, an American stock index representing the nation’s largest companies, at the peak of the bubble would not break even until six years later, while the Nasdaq did not recover its losses until well over a decade had passed.

Those on the cusp of retirement faced enormous implications as the crash wrecked their plans, according to research published in 2000 by the National Bureau of Economic Research, an American think tank. The number of US workers retiring was projected to have dropped by 3pc compared with a scenario in which markets didn’t crash.

“Retirement-age individuals who were still working watched their nest eggs shrink, and many have undoubtedly concluded that it would be prudent to continue the income from working until either the market recovers or they have had an opportunity to rebuild,” the report found.

Today, UK pensions are more heavily invested in the US than ever. Over just the past decade, defined contribution pensions have reduced their investment in UK equities from a third to just 6pc – and the gap has been filled by investment managers turning their eyes westward.

Around 45pc of the total assets held in British pension schemes are currently invested in US shares, according to investment bank Panmure Liberum. And AI represents an ever-growing share of the US stock market. The 10 largest companies in the US, all but two of which are dominated by the AI story, now account for 39pc of the total value of the American market, meaning an enormous amount of our pensions is tied up in the AI boom.

Since the launch of ChatGPT in 2022, investment in the new technology has exploded, but experts are warning history may be about to repeat itself – and the UK’s soon-to-be retirees are exposed.

In December, the Bank of England warned of a sharp correction in the value of major US tech companies, while the International Monetary Fund and the Organisation for Economic Co-operation and Development (OECD) have also issued recent warnings over an “AI bubble”.

Michael Burry, the American investor who predicted the global financial crisis of 2008, has placed a bet against major AI companies, purchasing financial products that pay out if the share prices of these firms fall – a mirror image of his first “big short” when he bet against the housing market after identifying the risk of subprime mortgages.

Further to fall

Heavy investment in the US market long looked like a safe bet. Over the past 10 years, the S&P 500 has grown more than 230pc – compare that to the 59pc offered by the FTSE 100 over the same period.

However, the US economy is buoyed by the so-called “Magnificent Seven” tech stocks that include AI chipmaker Nvidia, as well as Apple and Alphabet, Google’s parent company.

Anyone who invested in all seven of these tech stocks just five years ago could have more than doubled their money by 2025, based on their market performance in that time.

These seven stocks now represent around one third of the S&P 500’s total value, leading to concerns the market is overly concentrated in the fate of too few firms. In 2000, the top 10 stocks only comprised 23pc of the index.

Because these companies invest heavily in AI infrastructure and products, their valuations are soaring, but if the investment case of the new technology came into doubt, the negative effects could quickly spread.

After a decade of burgeoning tech growth, the market cap of Nasdaq firms is now almost five times higher than it was at the peak of the dot-com bubble – meaning it has considerably further to fall if things go wrong.

A bust would knock trillions of dollars off the value of the most vaunted American tech companies, hitting millions of savers.

Even without a crash, changing sentiment can wipe out hundreds of billions of dollars. Nvidia became the first company in the world to be valued at $5tn (£3.7tn) in October. This has already fallen to $4.4tn, shredding $600bn as a result of mild sentiment changes.

AI bubble could ‘destroy’ pensions

There are signs UK pension providers are already engaged in damage control to minimise their losses should things go wrong. Standard Life, which oversees a £32bn pension fund, has roughly 60pc invested in North America, but recently said it was reducing this.

Tom Hibbert, chief investment strategist at Canaccord Wealth, warns that workers considering retiring in the next five years should look carefully at how their pensions have been invested.

In the event of another market crash, there is a risk that, as in 2000, those with their money in US stocks and shares would be unable to stop working.

Hibbert warns that anyone in this generation on the cusp of retirement face their retirement portfolio being “destroyed” if they are overly exposed to equities during a crash – and there are currently around 4 million people in England and Wales set to hit the state pension age over the next five years, according to data from the Office for National Statistics.

“Getting your asset allocation right is key. If you are looking at retiring in the next five years you can’t tolerate a 50pc loss.”

This means it’s important to consider diversifying with other, potentially safer investments, such as government bonds or even valuable metals such as gold and silver.

Concerns over AI-linked companies are chiefly centred around the lifespan of computer chips produced by firms such as Nvidia, called graphics processing units (GPUs), that are needed to train the new software.

They require vast amounts of power and are stored in data centres. Over time their deterioration eats away at potential revenue. Some investors have begun to question whether these GPUs will last more than one or two years each, given they are already straining under the weight of enormous demand from the AI assistants they power.

The problem is, the high valuations placed on firms such as OpenAI, creator of ChatGPT, rely on these chips lasting five or six years.

“If you are a tech firm and you spend £100bn on GPUs and you depreciate them on your balance sheet over four years but they [actually] depreciate over two years, then you have effectively made up £50bn of earnings,” explained Hibbert.

If this shorter lifespan proves to be true, it could be the catalyst for a revaluation, or correction, of the market, wiping out imminent retirement plans.

Recommended

Don’t panic

This is concerning, but does not guarantee that a repeat of 2000 will happen, Hibbert said.

A key indicator of whether a company is overvalued is its price-to-earnings ratio, which shows how much its shares are selling for compared with how much money it is making.

Compared with Cisco, a digital infrastructure business that was briefly the world’s most valuable company shortly before the dot-com bubble burst, Nvidia is still far below the level of excessive market valuation seen in 2000. Nvidia’s share price is currently more than 45 times its earnings; Cisco’s was more than 200 times.

“I don’t think there is a huge risk. We’re in a very different place to where we were in the internet bubble,” Hibbert said. “The actual earnings are there, a lot of these companies are much more mature with very robust balance sheets.”

Even if the AI bubble were to burst, it might not do so until well after today’s older workers retire. Previous market bubbles have lasted years or even decades, meaning AI could continue to make money for investors for a long time to come.

Workers preparing to retire could try to minimise their exposure to companies linked to AI, but should do so in the knowledge that, given how these firms now dominate the world economy, their failure would spread far and wide. Protecting yourself entirely is very difficult.

Charles Hall, head of research at investment bank Peel Hunt, said it is still “too early” to call the growth of AI-related industries a bubble.

“The bubble has been talked about for some time but stocks are still pretty close to their all time highs.”

For in-depth analysis and expert opinion

Those who are concerned should check whether their pensions are linked to tracker funds, which tend to follow the direction of the largest stock indexes.

This will likely weight investments heavily in favour of the AI industry, given the tech-heavy US market comprises 65pc of the global stock index.

Younger workers will have a bigger share of their defined-contribution pension invested in AI-related tech companies than any demographic, but with retirement for this age group still decades away, their savings have more time to recover if the market goes edgeways edgeways.

Those with retirement on the imminent horizon, however, wouldn’t be so lucky.