Nvidia (NVDA 0.48%) has been a massive long-term winner from the AI megatrend, even if its stock has retreated a bit since early November. It's still up by more than 1,100% since heavy AI infrastructure investments started in 2023, and up by over 30% during 2025 alone.
The recent sell-off in the stock was driven by a combination of fears that the AI sector broadly is in a bubble and that Nvidia specifically is at risk of losing its dominance in the AI accelerator chip niche. While those fears are legitimate, I believe that Nvidia is just getting started. There are too many projections from the companies that are actually spending the money on AI infrastructure regarding the future of the industry, and those are the voices I trust most in this space.
With that in mind, I think investors should take advantage of Nvidia's now-lower stock price heading into 2026.
Image source: Getty Images.
Nvidia's GPUs are still the most popular AI computing units
Nvidia's graphics processing units (GPUs) have been the most popular AI computing stack since the artificial intelligence race began in 2023. The combination of its leading GPUs and best-in-class support software made Nvidia's stack the go-to choice for nearly every data center operator, and that hasn't changed.
Yet recently, news broke regarding a potential deal that challenged that notion: Meta Platforms (META 0.55%) may purchase a large quantity of tensor processing units (TPUs) from Alphabet (GOOG +0.34%) (GOOGL +0.20%) -- AI chips that it designed in collaboration with Broadcom (AVGO +0.88%). Meta is a huge Nvidia customer, so this turn of events is concerning. However, as Nvidia CEO Jensen Huang noted during its third-quarter earnings call, it is "sold out" of cloud GPUs. That context for the possible TPU deal could be key: Meta may have wanted to expand its cloud infrastructure buildout plans for 2026, found Nvidia unable to meet its increased demand, and turned to Alphabet as an alternative supplier.
NASDAQ: NVDA
Key Data Points
Only time will tell just how large this deal actually is, or if it will even happen at all. The news reports emphasized it was a potential purchase, not an actual order, and so far, Alphabet has deployed all of its TPUs in its own data centers. That said, I think more diversification in the AI chip industry is healthy, as it also keeps Nvidia from getting complacent in its leadership position.
The bubble conundrum
It's understandable why some investors are worried that the AI infrastructure industry could be repeating some of the patterns that occurred during the dot-com bubble. While I get the assertions of circular financing relating to the way OpenAI is funding its buildout based solely on its valuation, it's just one company in the industry. AI hyperscalers like Alphabet and Meta Platforms have real cash flows with which to fund their AI aspirations, and they have made it clear to investors that building out massive data centers to power these AI workloads is the only way to go.
As a result, there could be a slight AI bubble forming around one company (OpenAI), but the rest of the spending is legitimate. As a result, I think investors should still feel comfortable investing in Nvidia, as its future looks incredibly bright.
Nvidia can lose market share and still win for shareholders
Some estimates pegged Nvidia's share of the data center AI accelerator market at around 90% at its peak. Given how much money is pouring into this space and how large an opportunity there is, it was only a matter of time before other companies' competing hardware began to cut into that lead. However, the opportunity is so big that Nvidia can keep growing sales rapidly even as it loses some market share to rising competitors such as Alphabet.
Nvidia has projected that global data center capital expenditures will be around $600 billion in 2025. However, it expects that figure to rise to between $3 trillion and $4 trillion by 2030. That's a compound annual growth rate of 42% at the midpoint of that projection, which would be an incredible result.
Even if Nvidia's growth rate over the next five years is 30%, it would still be a wildly successful stock pick from here.
All of this will require the AI hyperscalers to continue devoting every penny of their cash flows toward AI buildouts, and for AI to generate a meaningful amount of money, either through cost savings or subscriptions. Time will tell whether that's how matters play out on the second point, but the AI hyperscalers have all announced record-breaking capital expenditures for 2026, supporting the first part of the requirement for another year.
In short, Nvidia can still be a highly profitable stock holding in the coming years even if it loses some market share. Investors should take its recent slight pullback as an opportunity to scoop up shares.
