Nvidia (NVDA +0.19%) has been a clear winner of the artificial intelligence (AI) revolution so far. The company has become the leading AI chip designer, holding about 80% of the market. Nvidia makes the world's most powerful graphics processing units (GPUs) -- these are the chips crucial for major AI tasks like the training and inferencing of models.
The biggest technology companies have flocked to Nvidia for these much-sought-after GPUs to power their AI projects, and that's resulted in double- and triple-digit revenue growth for Nvidia quarter after quarter. And this revenue has reached record levels, with Nvidia reporting more than $35 billion in the most recent period. On top of this, Nvidia is highly profitable on sales, with gross margin exceeding 70% on an ongoing basis.
But one particular risk has attracted investors' attention in recent times -- a risk that's intensifying and could potentially hurt revenue growth. It's linked to actions taken by some of Nvidia's customers. Are these tech giants about to say "checkmate" to Nvidia? Let's find out.
Image source: Getty Images.
Nvidia's most powerful chips
First, let's talk about Nvidia and its customers. Nvidia sells a variety of products and services to AI customers, but a point of focus has been on the company's GPUs. As mentioned, they're the most powerful -- and at $30,000 a pop, the most expensive -- out there right now. And Nvidia has pledged to update them on an annual basis, a move that could keep the chip designer ahead in terms of technology and quality of product.
Nvidia doesn't reveal the names of its biggest customers, but by considering comments from some of the world's major tech companies, we can identify them. Companies from Amazon to Meta Platforms and OpenAI have spoken of their use of Nvidia chips. For example, a year ago, Meta predicted that by the end of 2024, it would have 350,000 of Nvidia's H100 GPUs on board.
Another example: Oracle co-founder Larry Ellison said a few months ago that he and Tesla chief Elon Musk "begged" Nvidia for more chips -- this is understandable, considering demand for the company's latest Blackwell architecture and chip has surpassed supply.
So, it's clear that tech giants have been turning to Nvidia to power their projects since the start of the AI boom. Now, let's consider the move that could represent a risk for the chip leader. It has to do with competition, but not from rivals we would expect such as fellow chip designers Advanced Micro Devices or Intel. Instead, this competition comes from the research and development departments of Nvidia's customers.
Meta's "pursuing cost efficiencies"
Many have been producing their own chips, either as a way to reduce their dependence on Nvidia or to offer their customers access to lower-priced chips. Amazon has developed the Trainium line of chips and sells them to its cost-conscious customers through its Amazon Web Services business. Meta said during its recent earnings report that it's "pursuing cost efficiencies" by deploying its MTIA chip in ranking and recommendation inference workloads for ads and organic content.
And just recently, Reuters reported that OpenAI is finishing up designs for its own AI chip and expects to launch the manufacturing stage in the coming months.
Finally, Chinese start-up DeepSeek's report last month that it trained its model with Nvidia's lower-performance chips could reinforce the idea that companies may not always need to invest a fortune to get the job done.
If we consider these moves by tech giants, it may look as if they are about to say "checkmate" to Nvidia -- leaving the chip leader with slower revenue growth as these players turn to lower-cost chips, and in many cases, those they've developed in house.
NASDAQ: NVDA
Key Data Points
Is the chess game over?
But before we call this chess game over and declare Nvidia the loser, it's important to consider the whole picture. Yes, these Nvidia customers are expanding their chip options. But I don't necessarily think this will hurt growth at Nvidia. Here's why. First, tech giants are looking for excellence so clearly will seek the best -- and so far that's Nvidia -- to propel their AI programs. So, they may deploy Nvidia's latest GPUs for the most crucial of tasks and let some of their own chips play supporting roles. And it's in a customer's best interest to build on their current Nvidia system in order to reduce total cost of ownership over time.
It's also important to remember that Nvidia doesn't sell only GPUs but a whole ecosystem of products and services to advance AI projects. So, the company's revenue doesn't depend on GPUs alone. And as the AI revolution moves into the next stages, Nvidia's revenue opportunities actually could expand -- for example, more companies may turn to Nvidia's Enterprise software offerings.
All of this means that Nvidia has plenty of potential moves on the chess board, and ones that could ensure revenue growth and share performance as this AI boom continues.
