When it comes to training generative AI models, Nvidia's (NASDAQ: NVDA) graphics processing units (GPUs) are hailed as the gold standard among industry experts. That's not exactly a novel conclusion considering the semiconductor powerhouse has amassed an estimated 90% or more of the GPU market.
The more subtle idea here is how exactly Nvidia built such a gigantic lead over the competition. While it does not explicitly specify which companies buy its GPUs, it's widely speculated across Wall Street that cloud hyperscalers Microsoft, Alphabet, and Amazon, as well as Meta Platforms are among Nvidia's largest customers.
Where to invest $1,000 right now? Our analyst team just revealed what they believe are the 10 best stocks to buy right now. Continue »
Considering that these players are forecasting well in excess of $300 billion on AI infrastructure spending just this year, I wouldn't be shocked at all if these "Magnificent Seven" members are repeat customers of Nvidia.
But beyond the typical mega-cap AI names is another company quickly emerging as a top client. Let's explore how Elon Musk's new start-up, xAI, is deploying Nvidia's hardware and assess just how much it could be spending on the industry's best chip architecture.
Musk's xAI is a start-up building a large language model (LLM) called Grok, which is meant to compete with the likes of other popular LLMs such as ChatGPT, developed by OpenAI.
For the last year, xAI's primary focus was building a supercomputer to train its AI applications. The initial stage of development for the supercomputer, called Colossus, used 100,000 Nvidia GPUs. Shortly thereafter, Musk and his team scaled up its GPU cluster to 200,000 chips.
Image source: Getty Images.
A couple of months ago, Musk sat down with a team of developers and talked shop about how Grok is trained and what's next for xAI. Perhaps unsurprisingly, he doubled down on securing more GPUs.
At the time, he implied that the next training cluster would be five times larger than the current infrastructure. In other words, Colossus 2 would need 1 million chips (five times the 200,000 referenced above). He estimated the total cost for this project would fall between $25 billion and $30 billion.
However, according to a more recent report from Beth Kindig, the CEO and lead tech analyst at the I/O Fund, Colossus 2 may cost closer to $40 billion. Considering Nvidia's chips sell for between $30,000 and $40,000 on average, it's not unreasonable to think Musk's next big project could cost tens of billions of dollars -- as Kindig implies.
One of the main concerns surrounding an investment in Nvidia right now is the rise of custom silicon. Even though the hyperscalers all buy its GPUs today, each is also working on developing its own chips in-house. When you layer on top that Advanced Micro Devices (or AMD) is quickly gaining steam in the data center realm, there are legitimate reasons to be concerned about Nvidia's current growth trajectory.
NVDA Revenue Estimates for Next Fiscal Year data by YCharts; EPS = earnings per share.
Check out Wall Street's revenue and earnings forecasts above for Nvidia over the next couple of years. While the company's growth may begin to show some signs of deceleration, I'd say this is actually quite normal for a mature business, especially in the face of rising competition.
The bigger idea underscored by the financial profile above is that industry analysts are still forecasting growth for Nvidia over the next couple of years. In addition, if xAI does spend near the higher end of the estimated total cost I referenced above, this also showcases that the chipmaker is able to maintain some relative pricing power despite the intensifying competitive landscape.
I think these trends suggest that even if the cloud hyperscalers begin to diversify away from Nvidia, and AMD maintains its current momentum, it is still in a unique position, with emerging customers such as xAI willing to absorb demand from traditional customers.
Right now, Nvidia stock is trading near its lowest price-to-earnings (P/E) multiple in a year. I think concerns around tariffs, export controls in China, and rising competition have all made for a short-term souring on the semiconductor leader. But if the forecasts above and Musk's aggressive buying of Nvidia GPUs say anything, it's that the company shouldn't have much trouble fulfilling demand.
I think now is a great opportunity to take advantage of Nvidia's stock price action and prepare to hold for the long run.
Ever feel like you missed the boat in buying the most successful stocks? Then you’ll want to hear this.
On rare occasions, our expert team of analysts issues a “Double Down” stock recommendation for companies that they think are about to pop. If you’re worried you’ve already missed your chance to invest, now is the best time to buy before it’s too late. And the numbers speak for themselves:
Right now, we’re issuing “Double Down” alerts for three incredible companies, available when you join Stock Advisor, and there may not be another chance like this anytime soon.
*Stock Advisor returns as of May 12, 2025
John Mackey, former CEO of Whole Foods Market, an Amazon subsidiary, is a member of The Motley Fool’s board of directors. Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool’s 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. Adam Spatacco has positions in Alphabet, Amazon, Meta Platforms, Microsoft, and Nvidia. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Meta Platforms, Microsoft, and Nvidia. 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.