Even from Luxembourg, where investors track the AI boom with mounting curiosity, Nvidia’s recent record valuation has revived an old concern: is the sector running ahead of itself? Investment experts at Lombard Odier, Pictet, Liontrust, VettaFi and Quintet argue the fundamentals tell a very different story.
Nvidia, one of the key companies at the heart of the AI boom, last month announced that its Q3 revenue came to 57 billion dollars, up 66 percent from the year before. The chipmaker at the end of October attained a peak market valuation north of 5,000 billion dollars, giving up some of its recent gains since then. But the AI frenzy remains intense. To some, the momentum recalls the dotcom bubble at the turn of the century. Can the artificial intelligence train continue to steam ahead?
There are indeed some similarities between the dotcom bubble and AI excitement, said Daniele Antonucci, chief investment officer at Quintet Private Bank, speaking to Investment Officer. These include “a limited number of companies dominating global markets and heavy capital spending, but we believe there are many more significant differences,” he said. “First, the success of the dominant US tech companies has been mostly driven by strong earnings, which so far have been supporting valuations.”
“It is difficult to draw a parallel with previous bubbles where profitability was low, even non-existent, and debt was excessive. This is not the case with AI.”
Christopher Dembik, Pictet
Alphabet’s Q3 revenue, announced at the end of October, rose to 102 billion dollar, marking the company’s first ever 100 billion dollar quarter, CEO Sundar Pichal said during the quarterly earnings call.
No extremes
“Second, while valuations are high, they are still far from the extremes of the dotcom era,” noted Antonucci. “And third, unlike in the late 1990s, today’s so-called AI ‘hyperscalers’ – such as Amazon, Alphabet and Microsoft – are funding AI from cash flow rather than relying on debt.”
Christopher Dembik, senior investment adviser at Pictet Asset Management, also thinks that the current valuation of tech giants is not “excessive,” given their performance. Take, for instance, the example of Amazon.
“The company is a cash machine,” he told Investment Officer. “Its revenue was 638 billion dollars in 2024 and it generated a net margin of 10.5 percent.” What’s more, “Amazon has little debt and a lot of cash, 54 billion dollars at the end of 2024. Its price-earning ratio is around 30. This is not excessive. And, apart from a few exceptions like Tesla, all the major AI players are in a similar situation. It is difficult to draw a parallel with previous bubbles where profitability was low, even non-existent, and debt was excessive. This is not the case with AI.”
Hyperscalers boost investments
While there may be “some elements” of a bubble, it’s not on the point of exploding. said Luca Bindelli, head of investment strategy at Lombard Odier. “We continue to see an acceleration in investments; we know that the hyperscalers have started to announce bigger plans. These plans are now 32 percent more elevated compared to last year; at the start of the earnings season, it was around 20 percent-plus, so even there they have managed to surprise to the upside in terms of investments.”
“The price action is justified by the earnings fundamentals. This stands in sharp contrast to what we saw in the dotcom bubble period, where we saw irrational exuberance building relatively quickly, with earnings not justifying the price action.”
Luca Bindelli, Lombard Odier
Another way in which the dotcom bubble and AI boom differ comes down to fundamentals. “For now, the model is still able to deliver earnings growth that is consistent with the price action we’re seeing,” he said. “The price action is justified by the earnings fundamentals – and this stands in sharp contrast to what we saw in the dotcom bubble period, where we saw irrational exuberance building relatively quickly, with earnings not justifying the price action.”
Like Antonucci and Dembik, Bindelli added that “valuations are way more sustainable today than they were back then.” We’re not at dotcom levels. Indeed, we’re “far from it,” said Bindelli, but we’re still in a “historical context” with “relatively elevated valuations.”
There’s also some uncertainty about how monetisation will materialise in terms of revenues, he added. “I think this is really the game that’s being played at the moment. There’s no way out of further investing; the potential gains are going to be pretty sizable.” The hyperscalers are probably the best positioned for that, he noted.
Don’t just look at the models
It’s more than just the stalwart hyperscalers that are advancing in the AI space. The pace of innovation when it comes to AI is “astonishing,” said Storm Uru, co-lead fund manager and co-head of the global innovation team at Lionstrust Asset Management. But it’s not a bubble, he emphasised. Instead, a “changing of the guard” is underway, with industry incumbents ceding market share and new winners emerging.
For Axel Belorde, head of business development EMEA & Asia at the index provider VettaFi, whether or not the AI boom is a bubble depends on where one is investing.
“If you just buy foundational models, which are becoming more commoditised, if you just buy the compute capability in its current shape, and just big tech names, it’s not sufficient for diversified exposure to AI. So it is more prone to a bubble,” he told Investment Officer, highlighting the importance of investing in the “mid-layer” and the “semantic layer,” which aims to simplify interactions between data storage systems and nontechnical users.
More than the metaverse
That being said, massive amounts of money invested in a new technology brings to mind another hype from just a few years ago: the metaverse. Facebook parent company Meta reported that its Reality Labs segment saw an operating loss of 4.4 billion dollars for the third quarter of 2025, adding to billions of losses over the last few years.
Interest in the metaverse was simply “a combination of multiple themes,” Belorde said. “Some of them didn’t really have revenue associated with them, and we don’t like buying themes that don’t have revenue. And when you buy themes without revenue, you always have the risk that it fades.” And it turns out that despite some of these companies running valuable businesses, “people still care about real, tangible actions,” he added. “They don’t want to live in a dreamworld.”
It’s key to focus on “AI applications that are driving better outcomes for real companies and real use cases,” Belorde said. “If you have that approach, you avoid the buzzword-type situation where suddenly people think that everyone dreams of living in video game worlds and needs the metaverse to survive. Remember: it’s always about augmenting the human, not moving them into a virtual world or replacing them.”