It’s no understatement to say AI has reshaped global markets. Since ChatGPT launched in 2022, everyone from manufacturers to CEOs and investors has been hustling to keep up with the AI explosion.
The more fearful see AI as the harbinger of end times — and in some ways it is — but as investors worldwide have noticed, it can also be very profitable. And with that comes the rush to be first on the AI train.
“Because this is such a new industry, it is creating a lot of uncertainty, and that creates volatility,” says Mish-al Emeran, portfolio manager at Abax Investments. The market has responded by doing what it tends to do: grouping things into black and white boxes, while talking up the idea of “winners” and “losers”.
“But it’s such a nascent industry and there’s so much that still needs to be developed,” Emeran warns. Indeed, defining surefire winners and losers is an imperfect science — but one the market insists on experimenting in.
Some emperors with no clothes
There are, of course, obvious winners, says Emeran.
“The tech businesses, the hyperscalers, the Magnificent Seven, a couple of Chinese stocks that are into the production of chips.”
The Magnificent Seven are chip company Nvidia (whose shares have risen 1,369% over the past five years), Google holding company Alphabet (up 208%), Facebook owner Meta (up 156%), Apple (up 26%), Tesla (up 83%), Microsoft (up 74%) and Amazon (up 38%). Collectively, their returns have risen at more than double the rest of the S&P 500 over that time.
Likewise, the Semiconductor Manufacturing International Corporation has risen more than 170% in China.
Beyond that, companies involved in the physical build-out of AI infrastructure are also considered winners, says James Bennett, global equity analyst at Anchor Capital.
These include companies supplying the energy and cooling systems for data centres, such as Vertiv Holdings (up 810% over five years) and Bloom Energy (up 550%). Even older companies like Dell Technologies (up 730% over five years) are considered AI winners for their part in making physical hardware for AI.
Of course, it’s impossible to mention the AI darlings without talking about Anthropic. Its imminent public listing, aiming for a value of between $1.5-trillion and $2-trillion, may be the largest in history, eclipsing SpaceX’s $1.77-trillion.
But not all assumed winners are crowned — and not all investors are drinking the Anthropic Kool-Aid.
“I’m a huge AI fan and we own almost all of the major players, but I’m sceptical of Anthropic,” says Vestact CEO Paul Theron. “Mostly because I don’t trust [CEO] Dario Amodei to focus adequately on creating long-term shareholder value. He is a talented biophysicist and a good fundraiser, but he’s not a suitable person to be the CEO of a major corporation.”
In September, Amodei published a 3,800-word essay warning that AI is advancing at an “exponential” rate and should slow down for humanity’s sake.
This rankles Theron. “Can you imagine Henry Ford saying: ‘My cars are dangerous and could cause accidents, maybe you shouldn’t buy one’? Or Steve Jobs saying: ‘These smartphones could kill your attention span and turn you into a zombie, so we are going to slow down their development’?”
Meanwhile, a healthy dose of scepticism is good.
Bennett says the above-mentioned companies are all seen as “clear winners in the short to medium term”.
The concern for investors, though, is what analysts fret may be an “earnings cliff” in the long term. Profits from AI may slow sharply, even though the cost of building out the systems will only keep rising.
This could ricochet down the chain — if demand for chips and cooling infrastructure wanes, the “winners” of the first wave of AI firms may become the losers of the next.
Collateral damage
Perhaps more interesting than the winners is the predicament of the “losers” — businesses that, in most cases, aren’t even that bad.
“When the market is so preoccupied by the theme of the day, good companies, through no fault of their own, tend to get overlooked, and you see their earnings multiples trickling down over time,” Bennett says.
It’s clear, however, that some companies will be destroyed by AI — especially those that refuse to take it seriously.
“Kodak was one of the big players in the camera market and got wiped out because it didn’t want to change to digital cameras,” says Zwelakhe Mnguni, chief investment officer of Benguela Global Fund Managers. “That kind of complacency can create problems and then the loss of value is permanent.”
It’s the same with video rental company Blockbuster, which had 9,000 stores at one point and famously passed up the chance to buy Netflix for $50m in 2000. Blockbuster filed for bankruptcy in 2010; Netflix is now worth $254bn.
Still, many of the companies being trashed by the market are actually still quality investments, analysts say.
“Last year, everybody said OpenAI was going to wipe out [Alphabet], which really damaged the share price. But by the end of the year, Alphabet had been the best-performing big tech share of 2025,” says Bennett.
It’s the same with the major cybersecurity companies like Palo Alto Networks, CrowdStrike and Fortinet. Bennett says by February, their stock prices had fallen more than 20%, but since then have bounced back and are up more than 100% this year.
As Mnguni says: “In many cases, the market priced a certainty where there was only a risk.”
Many of the companies seen as “losers” have actually been able to defy the curse by integrating AI into their businesses.
Mnguni cites credit reporting agency Experian, which got beaten down “quite aggressively” by the market but “should benefit from using AI because its analytics engines should be even better now for the underwriting of credit”.
In Experian’s case, he believes the market overreacted, especially given that the company had solid revenue growth last year with a 28.6% pre-tax profit margin to boot.
Which isn’t to say that some companies, like Adobe, won’t be in genuine trouble. But more might have been unfairly undervalued in the rush for the “winners”.
In some cases, the market will get it right; in others, not, Bennett says. “We’re only going to find out in three to five years whether all these concerns are justified, but there will be some companies that the market currently perceives to be AI losers that will turn out to not be.”
Picking the winners
It’s a minefield: while some stocks remain undervalued, others are overvalued.
Here, Mnguni points to companies producing semiconductor components for data centres. “They were always oversupplied to some extent, and they got rescued by this AI build boom,” he says.
In this frenzy, companies like Diodes have exploded in size in the past five years but remain vulnerable to becoming irrelevant.
Bloom Energy has also been dangerously overinflated.
“You might find that the business model is good, but the valuation is problematic,” Mnguni says. Its p:e is 137, a “ridiculous” multiple that seems ripe for a crash. Investors who paid a pretty penny for Bloom shares will be disappointed to find they have tumbled 40% from their high in June.
The whims of the market can crater companies once deemed “winners”.
AI infrastructure company Cerebras Systems’ shares dropped 46% between September and October, ostensibly because of news headlines asking pointed questions about its technology, rather than any change in revenue.
So how do you pick the long-term winners?
There’s no easy way, the experts say. Analysts recommend diversifying to avoid heavy exposure to any one theme — “with 49% of the S&P 500 in one theme, an index investor is already making a concentrated bet”, Mnguni says.
Of his strategy, Emeran says: “We always try to bring it back to the underlying drivers of valuation. What impact will [AI] have on cash flows and future growth, and what is it going to do for risk?”
And most of all, pay attention to the companies the market has forgotten — they’ll get their day in the sun, Bennett says.
“When the hot trade of the day starts to come off the boil, the good-quality companies that have been pushed down to low earnings multiples will start to recover,” he says.