Artificial intelligence (AI) shares have been the talk of the markets in 2025. You'd have to have been stranded on a desert island to have missed all the headlines.
And AI has certainly proven itself in some fairly interesting use cases. Not to mention, it's potential impact to everything from work productivity, health, and problem solving.
But with a big crowd comes lots of noise. And boy, has it been noisy for the basket of AI shares trading on global stock exchanges.
But are they worth the hype? Leading hedge fund Man Group, situated in the UK, has run its numbers, drawing parallels of today's AI hype to the infamous "DotCom bubble" of almost 25 years ago.
Let's take a closer look.
Image source: Getty Images
Are AI shares the same as 'DotCom' companies?
Man Group is one of the oldest hedge funds in existence. It manages a colossal amount of money for global investors, and has access to the latest trends in technology. AI included.
The firm points out that, much like the DotCom bubble two decades ago, today's valuations of AI shares are heavily concentrated among a few tech giants.
Some of these include Google parent Alphabet Inc. (NASDAQ: GOOG), Meta Platforms (NASDAQ: META), formerly Facebook, and global chip powerhouse NVIDIA Corp (NASDAQ: NVDA).
For reference, the DotCom bubble refers to the large run up of internet companies in the late 1990's and early 2000's. Many of these were unprofitable, despite trading at sky-high valuations in the market.
Naturally, this led to a sharp crash in global markets when the bubble burst, after many investors got over the hype.
But Man Group sees plenty of similarities between previous stock market bubbles, and now, with AI shares.
The DotCom parallel is hackneyed, but clichés are useful sometimes, and this one might be all we've got. There have only been two other US equity bubbles in the last 100 years (the 'roaring twenties' and the 'nifty fifty'), and both were much more broadly dispersed, covering sectors as variant as railroads, consumer goods, automobiles and financials, as well as nascent technologies.
The DotCom bubble was different in terms of its concentration of performance around internet innovation…
…And, to return to the original point, maybe AI isn't that difficult after all. Really helpful, but widely available, and not conducive to the creation of a new corporate aristocracy. In that eventuality, these multiples could end up as albatrosses.
All negative? Not so
Man is mostly talking about the US stock indexes, but there are some parallels with the S&P/ASX 200 Index (ASX: XJO) at the time of writing. It too has a growing influence of AI.
And while the firm points to the risks in potentially over-hyping AI shares, it's clearly not all negative.
In a report from March this year, global consulting giant McKinsey & Co, noted that companies are making "organisational changes" to generate value from AI.
It says that companies "with at least $500 million in annual revenue" are embracing AI faster than their smaller counterparts. This may or may not be positive for companies providing this tech.
But the big difference between the current excitement, and the DotCom era (and indeed every other bubble) is concentration.
The space continues to be dominated by one name: OpenAI's ChatGPT platform. Per Man Group's analysis:
OpenAI has 10 million individual paying customers, shelling out US$20 a month, and a further one million business clients on US$2,000 a month. So perhaps in the region of US$26 billion annual revenue.
Nothing artificial about the size of those numbers, that's for sure.
Plus, despite the cautious tone, some AI shares are reporting chunky productivity gains, according to Man Group.
Some are reporting 16% annual sales growth per employee, it says. Sales per employee is measured by taking a company's total revenues divided by total number of employees, and is a measure of productivity.
Foolish takeaway
Man Group has painted a mixed view on the state of AI shares. But the hedge fund also warns of valuation risks: buying great businesses at excessive valuations often leads to losses, as history showed with tech stocks in 2000.
Meaning, that if AI shares fail to deliver the extraordinary growth baked into their prices, investors could face significant headwinds if their market prices respond negatively to this.
And whist no one has a crystal ball, the lessons of the past do certainly ring true.
On that note, it's always wise to quote none other than Warren Buffett in these scenarios: "It's far better to buy a wonderful company at a fair price than a fair company at a wonderful price".
Sage advice in the current market.
