Nvidia’s 24.5x P/E Multiple Looks Like a Market Mistake — and the Value Argument Is Only Getting Louder the Longer Shares Drift Sideways

Even if the latest surge proves unsustainable and shares of Nvidia (NASDAQ:NVDA | NVDA Price Prediction) are destined to fly under $200 per share for another quarter or even another year, it’s becoming harder to argue against the value case, even if you think the margins will compress by a bit and the growth takes a few backward steps. At this juncture, the bull case, which might be more realistic than investors think, doesn’t seem to be taken into consideration.

With CEO Jensen Huang recently remarking that the firm has zero market share in China, I do think a bright light has been shone on what Nvidia could become if hurdles were cleared and a bull-case scenario is allowed to pan out.

In any case, the stock is starting to get too cheap again, now going for 24.5 times forward price-to-earnings (P/E), making it quite a bit cheaper than some of the Magnificent Seven names with growth profiles that aren’t nearly so magnificent.

Competition is getting tougher. But Nvidia still has legs

Of course, competition only gets tougher from here as more firms look to commercialize their AI chip offerings. As Nvidia looks to reinvest its profits into some incredibly compelling firms across the AI stack, though, I’d be more inclined to view Nvidia as becoming more of an all-around AI bet rather than just a seller of GPUs. 

While it’s harder to come by a bear beyond the legendary Dr. Michael Burry of The Big Short fame, I do think investors should consider the potential headwinds in play as well. After all, a high-growth stock with a valuation multiple that’s too good to be true may very well come with some strings attached. In the case of Nvidia, though, I do think it’s priced with such headwinds in mind and not with the possibility of maintaining its incredible earnings growth engine.

At less than 25 times forward P/E, I think betting against the name could be a dangerously risky move, given the shares don’t seem priced for perfection, but perhaps priced with a looming disappointment in mind. In any case, the GPU trade was so yester-year.

From GPUs to ASICs and training to inference

These days, it’s more about custom silicon and the hyperscalers that are innovating on with their own ASICs (application-specific integrated circuits). Also, investor interest seems to similarly be gravitating from training to inference, especially in the face of a rise in agentic AI.

With a software-esque gross margin north of 70% (Nvidia does pair its great software with chips, which is a part of what makes the Nvidia ecosystem so great) and a market lead that is its to lose, I can see why the skeptics aren’t biting on the multiple, no matter how low it gets.

At the end of the day, there’s no telling how long the multiple can go. When it comes to the cyclical stocks, perhaps the dirt-cheap-looking multiples are more of a danger sign and a coming peak than a sign of actual undervaluation. Given the AI revolution might be more of a supercycle than a normal cycle, though, investors might be mistaken to dismiss the ever-compressing P/E as a mirage, value trap, or anything similar.

The bottom line

As shares drift sideways and Nvidia keeps putting up great quarters, the stock stands to only get cheaper and cheaper. The only big question is at what point does it become ridiculous to keep passing up on the name? Has such a point already been passed? Possibly. Either way, there’s no shortage of smart money bulls that are willing to step in as buyers at these prices.