Nvidia (NASDAQ: NVDA) supplies the most advanced graphics processing units (GPUs) for developing artificial intelligence (AI) models. Many of the world’s largest technology companies are spending boatloads of money to fill their data centers with those chips as they jostle for leadership in the AI race.
As a result, Nvidia’s data center revenue generated triple-digit percentage growth in each of the last six quarters. But some Wall Street analysts are questioning how long the AI spending boom can last.
Start Your Mornings Smarter! Wake up with Breakfast news in your inbox every market day.
During an earnings call with investors for the fiscal 2025 third quarter (ended Oct. 27), Nvidia CEO Jensen Huang aimed to put those concerns to bed. He also offered some insights into the demand picture for its new Blackwell GPUs, and here’s why investors should be very excited.
Developing AI isn’t cheap. A single data center GPU can cost up to $40,000, and the most advanced AI models require tens of thousands of them to deliver the appropriate amount of computing power. Nvidia’s H100 and H200 GPUs have been the go-to choice for AI development over the past year. They use the company’s Hopper architecture, which was the gold standard in performance and energy efficiency
But now, Nvidia’s new Blackwell architecture offers a major leap forward on both fronts. A Blackwell-based GB200 NVL72 system, for example, can perform AI inference at 30 times the speed of the equivalent H100 system, while providing similar improvement in energy efficiency. That translates into substantial cost savings for data center operators like Microsoft and Amazon, and the AI developers who rent computing power from them.
A single GB200 GPU within the NVL72 system sells for about $83,333, which is double the price of the H100 when it first came out. But considering the 30-fold increase in AI inference performance and comparable improvement in energy efficiency, AI developers are coming out way ahead even if they are paying double for these new chips.
Simply put, Blackwell is going to make the most advanced AI models financially accessible to a wider variety of businesses and developers.
Nvidia generated $30.8 billion in data center revenue during the fiscal 2025 third quarter, which was a 112% increase from the year-ago period. The company shipped only 13,000 sample Blackwell GPUs to customers, so the new chips weren’t a big contributor, but sales are expected to ramp up significantly from here.
Microsoft is reportedly the biggest Blackwell customer so far. We don’t know the size of its orders, but we do know the company allocated $20 billion to capital expenditures (capex) during its fiscal 2025 first quarter (ended Sept. 30), which followed $55.7 billion in capex spending throughout fiscal 2024. Most of that money went toward AI data center infrastructure and chips.
Amazon is another top Nvidia customer. Its AI capex spending is on track to hit $75 billion in calendar 2024, and then there is Meta Platforms, which will spend up to $40 billion this year. Oracle will also provide investors with a capex update in December, but we already know the company plans to build clusters with 131,000 Blackwell GPUs.
However, that kind of spending can’t continue in perpetuity, and some analysts are already expressing caution. While Goldman Sachs is bullish on AI, the investment bank concedes that a killer AI software app is yet to emerge to justify the substantial investments these tech companies are making. Plus, Goldman says if this AI capex only results in customer service chatbots and code generators, then the tech sector is massively overspending.
In the end, investors will eventually demand a return on all of that spending.
Despite any concerns, Nvidia investors probably shouldn’t feel nervous at this stage. In his third-quarter conference call with investors, Huang described Blackwell demand as “staggering.” He originally said the new chips will contribute “several billion dollars” in revenue in the fourth and final quarter of fiscal 2025, but he said the company is likely to exceed that estimate (although he didn’t provide a firm dollar figure).
When asked about a digestion period following all the AI spending so far, Huang said he doesn’t expect any slowing in spending until $1 trillion worth of existing data centers are modernized with new GPUs over the next several years. That means Nvidia’s sales could steadily grow until 2030 before data center operators take a pause.
And not all analysts are worried about a potential slowdown. Morgan Stanley estimates Microsoft, Amazon, Meta Platforms, and Alphabet will spend a combined $300 billion on AI infrastructure in 2025 alone. The investment bank predicts Nvidia is on track to ship up to 300,000 Blackwell GPUs in the final three months of 2024, followed by as many as 800,000 in the first quarter of 2025.
Therefore, investors who already own Nvidia could do well to hold on for the long run. And it’s probably not too late for new investors to buy into this incredible story, either.
Ever feel like you missed the boat in buying the most successful stocks? Then youโll want to hear this.
On rare occasions, our expert team of analysts issues a โDouble Downโ stock recommendation for companies that they think are about to pop. If youโre worried youโve already missed your chance to invest, now is the best time to buy before itโs too late. And the numbers speak for themselves:
Nvidia: if you invested $1,000 when we doubled down in 2009, youโd have $368,053!*
Apple: if you invested $1,000 when we doubled down in 2008, youโd have $43,533!*
Netflix: if you invested $1,000 when we doubled down in 2004, youโd have $484,170!*
Right now, weโre issuing โDouble Downโ alerts for three incredible companies, and there may not be another chance like this anytime soon.
John Mackey, former CEO of Whole Foods Market, an Amazon subsidiary, is a member of The Motley Foolโs board of directors. Randi Zuckerberg, a former director of market development and spokeswoman for Facebook and sister to Meta Platforms CEO Mark Zuckerberg, is a member of The Motley Fool’s board of directors. Suzanne Frey, an executive at Alphabet, is a member of The Motley Foolโs board of directors. Anthony Di Pizio has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Alphabet, Amazon, Goldman Sachs Group, Meta Platforms, Microsoft, Nvidia, and Oracle. The Motley Fool recommends the following options: long January 2026 $395 calls on Microsoft and short January 2026 $405 calls on Microsoft. The Motley Fool has a disclosure policy.