As AI bubble fears grow, shares of GPU titan Nvidia (NASDAQ:NVDA) seem like a top candidate for some profit-taking, especially as things get a bit choppier and the momentum begins to taper further. Although it may not feel like it right now, Nvidia is still a major multi-bagger in the past two years, with over 240% in gains in the books.
Of course, that two-year gain is front-loaded, to say the least. Year to date, the stock is up just shy of 24%, and many investors who’ve gotten in more recently might be wondering if there’s another leg higher in 2026, or if the best days really are over.
Undoubtedly, Nvidia may still be miles ahead of its top rival in GPUs, Advanced Micro Devices (NASDAQ:AMD), but you wouldn’t know it by comparing the past-year performance of the two stocks.
Nvidia stock has already corrected amid AI bubble fears
Either way, growing distaste for AI stocks and Dr. Michael Burry’s very public bearish bets on Nvidia have put Nvidia in a rather difficult position. Shares of Nvidia are now sitting down more than 17% from recent highs, and while Burry might not be big in the money yet with his put options, it might not take long before he is, especially if Nvidia finds itself in a bear market by the time Santa Claus rally season and the new year ring in.
Of course, there’s a huge chance that big AI spenders have gotten ahead of their skis, and, as a result, might need to cut their spend in the new year, perhaps deeper than expected. Undoubtedly, AI demand is showing no signs of slowing down, but eventually, scaling laws might shift, and adding more GPUs to the mix might lead to diminishing returns. In the meantime, though, the hyperscalers seem more willing to keep on “flooring it” going into the new year.
More of the same in 2026? Why scaling up and backing up the truck on GPUs might still be the theme
Google DeepMind CEO Demis Hassabis believes that AI scaling “must be pushed to the maximum.” Of course, there are other aspects, including new approaches, architectures, and more research that ought to be considered to keep the AI ball rolling as fast as it can.
In any case, there still might be more low-hanging fruit that hyperscaling may be able to grab. And I do think it just makes sense to ensure all the fruit is grabbed before stopping efforts to pick them. Worst case, the AI spenders can adjust their spend accordingly, entering another year of efficiency, so to speak. The last year of efficiency surely powered the hyperscalers to euphoric gains.
Given the bad taste that AI spending has given investors, I’d be inclined to bet that another course correction would also cause an upward correction in shares of the heavier AI spenders. I view it as an option that big tech will exercise very carefully. While we’re at these heights, I’d argue that continuing to scale up is the way to go, at least until we do reach AI advancement returns that have rapidly diminished. And for the time being, Nvidia stock is positioned to keep winning big, as frontier training remains at the top of mind.
Sure, it’s nice to get out of the AI-spend-heavy trade well ahead of time, but doing so also accompanies the risk of missing out on additional gains. And if we are in the middle innings of the ballgame, a premature exit from the trade might lead some to buy back into the top AI plays at higher prices and closer to the peak of this AI boom.
Nvidia stock is a massive bargain if you don’t see an AI bubble
Given all the uncertainty, I think sticking with Nvidia makes sense, especially as the valuation begins to come in again. If we are assuming that Nvidia’s next generation of GPUs will keep selling, the 23.4 times forward price-to-earnings (P/E) multiple on Nvidia stock seems way too low.
Of course, if more AI compute leads to diminishing improvements for AI models, perhaps GPU sales will drop off very sharply, and Nvidia stock could be one of the biggest losers as the AI trade (or bubble) goes bust.
So, either Nvidia stock is ridiculously cheap or is obscenely expensive, depending on your views of the AI trade and which inning you see the boom in. My guess is we’re not yet at the peak, and that means another year or more of big gains in the cards for the GPU kingpin.