The debate over whether artificial intelligence (AI) stocks are in a bubble has intensified this year. Massive spending on AI infrastructure, soaring valuations across parts of the semiconductor sector, and comparisons with the dot-com era have left investors asking whether the enthusiasm reflects durable growth or whether expectations have simply run too far ahead of fundamentals.
DBS Group Chief Investment Officer Hou Wey Fook recently weighed in on that debate, arguing that AI stocks remain far from bubble territory. Interestingly, his case centers on Nvidia (NVDA), the company that has become perhaps the clearest symbol of the AI boom. Given Nvidia’s central role in the buildout of AI infrastructure, Wey Fook believes the chipmaker offers an important clue about where the broader AI trade stands today.
His argument is compelling in some respects, but using one company—even the industry leader—to draw conclusions about an entire sector deserves a closer examination. So, does Nvidia really prove that AI stocks aren’t in a bubble, or is the picture more complicated? Let’s take a closer look.
DBS Says Nvidia’s Valuation Shows AI Stocks Aren’t in a Bubble
The debate over whether AI stocks are in a bubble just got a fresh take from DBS Group Chief Investment Officer Hou Wey Fook. His argument was straightforward: he pointed to Nvidia’s price-to-earnings (P/E) multiple and projected earnings growth for next year as evidence that AI stocks remain far from bubble territory.
Nvidia currently trades at 15.03 times projected earnings for fiscal 2028, which begins in February 2027. In its second-quarter earnings report in August, the company projected FY28 sales growth of 70% year-over-year (YoY), well above the 45% analysts had expected. Analysts now forecast roughly the same growth rate for earnings. Wey Fook said that marks a stark contrast with Cisco near the peak of the dot-com boom, when the stock traded at about 100 times earnings and its valuation depended much more heavily on future expectations being met.
In a Bloomberg TV interview, Wey Fook argued that it is difficult to call AI a bubble when its “poster child” trades at a mid-teens earnings multiple, adding that semiconductors and AI still have meaningful tailwinds.
Well, I think Wey Fook has a strong argument when it comes specifically to Nvidia, but the conclusion becomes less convincing when it is extended to AI stocks as a whole. Allow me to explain this in more detail.
Jensen Huang Calls Nvidia a “Growth Value Stock”—The Numbers Back Him Up
The core of Wey Fook’s argument is particularly compelling for Nvidia, whose valuation has compressed largely because earnings estimates have surged while the stock has spent much of the year trading sideways. Nvidia’s forward P/E multiple fell to near its lowest level in more than a decade in September, even as the company is expected to deliver more than 90% YoY growth in both revenue and earnings per share (EPS) this fiscal year.
The gap between Nvidia’s robust fundamentals and relatively modest valuation led CEO Jensen Huang to call the company “the world’s first and only growth value stock.” Speaking at a Goldman Sachs technology conference in September, Huang described Nvidia as “incredibly misunderstood,” adding that “not only are we growing, we’re also capturing more at the same time.” Indeed, a 15.03x earnings multiple would be more typical of a slow-growing, mature semiconductor company than one still expected to deliver strong growth.
Of course, Nvidia has faced concerns over its gross margin, which is expected to narrow in the coming quarters largely because of rising costs for key components, as well as competition from hyperscalers’ custom processors and worries about circular financing. Those issues restrained the stock. However, during the Q2 earnings call and in subsequent remarks, management has repeatedly sought to downplay most of them.
And more importantly for NVDA and other AI stocks, there have been no signs of a slowdown in AI demand. Nvidia AI server partner Foxconn reported robust September-quarter revenue on Monday and said its “AI-related operations” were expected to continue growing in the fourth quarter. Separately, AMD (AMD) CEO Lisa Su said on Tuesday that chip demand is likely to remain “very high” for the next several years.
Taken together, these factors make it difficult to argue that NVDA itself is trading in bubble territory. The stock’s valuation remains heavily restrained given its growth outlook, while demand for AI infrastructure remains strong.
Why Nvidia’s P/E Is a Poor Gauge of the Broader AI Trade
The broader AI trade, however, presents a more complicated picture, and Wey Fook’s argument becomes less convincing when applied beyond Nvidia. To be clear, I don’t believe any part of the AI trade—or the AI trade as a whole—is currently in a bubble. My point is simply that valuations vary widely across the group, making Nvidia’s multiple a poor standalone gauge of where the broader AI trade really stands.
Let’s take a look at the hottest part of the AI trade—memory chipmakers, which trade at very low P/E multiples because of the historically cyclical nature of their businesses. Micron (MU), for instance, currently trades at a forward P/E multiple of just 5.94x. Even if that multiple were to triple to around 18x, it would still remain well below Cisco’s (CSCO) valuation before the dot-com crash. That highlights one problem with using Nvidia as a proxy for the broader AI trade: some segments structurally command much lower multiples because investors apply a larger discount to cyclical earnings.
Conversely, some prominent chipmakers, including Marvell Technology (MRVL), Intel (INTC), and Advanced Micro Devices, trade at forward P/E multiples ranging from 68.10x to 85.60x. For some investors, those valuations may look far more bubble-like, as they approach Cisco’s multiple near the peak of the dot-com boom. That highlights the opposite issue: some AI-linked stocks trade at far richer multiples than Nvidia, in part because their shares have posted triple-digit gains this year.
With that, the wide dispersion in valuations makes Nvidia’s P/E a poor standalone gauge of whether the broader AI trade is in bubble territory. Instead, I believe investors should focus on the strength of AI demand and, more importantly, whether the AI boom is generating strong returns not only for chipmakers but also for their customers. And the upcoming earnings season will provide fresh insight into both.
On the date of publication, Oleksandr Pylypenko had a position in: NVDA . All information and data in this article is solely for informational purposes. For more information please view the Barchart Disclosure Policy here.