Founded in 1993, The Motley Fool is a financial services company dedicated to making the world smarter, happier, and richer. The Motley Fool reaches millions of people every month through our premium investing solutions, free guidance and market analysis on Fool.com, personal finance education, top-rated podcasts, and non-profit The Motley Fool Foundation.
Founded in 1993, The Motley Fool is a financial services company dedicated to making the world smarter, happier, and richer. The Motley Fool reaches millions of people every month through our premium investing solutions, free guidance and market analysis on Fool.com, personal finance education, top-rated podcasts, and non-profit The Motley Fool Foundation.
Artificial intelligence (AI) is not a topic that will end this year. There are still huge gains to be made in this sector, so investors aren’t shying away from it just because it has had two strong years. Nvidia (NVDA 4. 56%) is a stock that embodies investment in AI. A key point about Nvidia that sets it apart from other AI investments is that it makes a lot of money from all of the big tech companies’ AI investments.
Hans Mosesmann of Rosenblatt Securities has a top price target of $220, in line with Nvidia’s share. This means there will be a 67% increase over the next year. That’s a huge gain for a name the size of Nvidia, but it’s entirely imaginable if certain things happen.
As mentioned above, one key thing that sets Nvidia apart from other possible AI options is that it makes money from AI. While AI hyperscalers are investing hefty sums to expand their computing power to exercise AI models, in the hope that those investments will eventually pay off. Nvidia is keeping some of that money.
Its graphics processing units (GPUs) and software package that powers them are best-in-class and have become the go-to choice for any company looking to train an AI model. GPUs have a key ability to compute in parallel, which means they can process multiple calculations simultaneously. Furthermore, GPUs can be grouped in clusters to multiply this effect. With some of the largest AI companies building servers with thousands of GPUs in them, it’s evident how quickly AI models can be trained.
However, that’s just with the old-generation models. While Nvidia’s legacy Hopper architecture GPUs provided impressive performance, its next-generation Blackwell GPU architecture provides huge performance gains. Blackwell can train AI models at four times the speed of its predecessor, making it a key upgrade for those looking for the best computing performance possible.
Another key factor is that GPUs don’t have the longest lifespans in data centers. Because they are used nearly constantly, they burn out quickly. According to an unnamed Alphabet (NASDAQ: GOOG) (NASDAQ: GOOGL) data center specialist, this can cause the lifespan of a GPU to last from one to three years. As a result, many of the GPUs purchased over the past few years may need to be replaced. Throw in the fact that AI demand is still growing and that these companies are still building out their computing capacity, and it’s clear how Nvidia can keep growing.
But Nvidia isn’t without its challenges.
A couple of headwinds popping up that could cause Nvidia some trouble in 2025 are a greater push for using CPUs for AI inference and custom AI accelerators. While Nvidia’s GPUs are considered some of the best available, they are power-hungry and aren’t as efficient as a CPU for simple tasks like AI inference. Inference occurs when a pre-trained model is used, so using a powerful, energy-hungry GPU doesn’t make sense when a cheaper, more efficient CPU could be used.
Additionally, many of Nvidia’s biggest users use traditional AI exercise accelerators. These chips would allow them to overtake Nvidia, as they would not want to pay a middleman to get the computing power provided through GPUs. However, those traditional AI accelerators require workloads to be securely configured for optimal use. If a general AI design is being developed or developers are simply testing certain exercise techniques, these traditional accelerators may not be an effective way to exercise AI styles.
While these are clear headwinds for Nvidia, I believe they are relatively minor and will only affect a portion of Nvidia’s business.
So, how can Nvidia’s stock soar 67% in 2025? Simple. It does what it’s expected to and gives a solid outlook for next year.
Right now, Nvidia trades for 52 times trailing earnings, which is near the cheapest level it has traded at over the past two years.
As a result, I don’t think investors can call Nvidia “overvalued,” especially when other tech corporations like Apple and Amazon are trading at 38 and 47 times earnings, respectively, despite much slower growth.
Nvidia is expected to increase its cash by up to 52% in fiscal 2026 (ending in January 2026), so if it maintains its profit margins and its valuation rises a bit, Nvidia’s inventories will increase in that 67% range. As a final starting point, there will have to be indications that 2026 (Nvidia’s 2027 monetary year) will also be strong; otherwise, inventory could simply sink in anticipation of a slow year.
We’re at a point where Nvidia’s expectations don’t outweigh the current business. Although some are selling Nvidia stock to take profits, there’s still plenty of upside available, and I think Nvidia is a great buy after its latest pullback.
Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool’s board of directors. John Mackey, former CEO of Whole Foods Market, an Amazon subsidiary, is a member of The Motley Fool’s board of directors. Keithen Drury has positions in Alphabet, Amazon, and Nvidia. The Motley Fool has positions in and recommends Alphabet, Amazon, Apple, and Nvidia. The Motley Fool has a disclosure policy.
Market data powered by Xignite and Polygon.io.