It’s safe to say Nvidia (NASDAQ:NVDA) has done the impossible.
The company went from a $2 trillion market capitalization to $3 trillion in just 96 days, a little over a financial quarter. For reference, Apple (NASDAQ:AAPL) took nearly three years from 2020 to 2023 to do the same.
The cause for this explosive growth? Insatiable revenue as a result of aggressive graphics processing unit (GPU) sales. Of course, for legal reasons, Nvidia doesn’t openly disclose who is buying the majority of these GPUs. But speculation about Nvidia’s biggest mystery customer could be valuable. It can give clues about who is doing the most with artificial intelligence (AI). After all, the industry requires nearly endless amounts of computing power to sustain its projects and growth.
Thankfully, Bloomberg has reported on Nvidia’s biggest customers in 2023, and the results are not surprising. Coming in at 15% of Nvidia’s GPU sales, was Microsoft (NASDAQ:MSFT) followed by Meta Platforms (NASDAQ:META) and Amazon (NASDAQ:AMZN) at 13% and 6% respectively. Alphabet (NASDAQ:GOOG, NASDAQ:GOOGL) came in at 6% as well, totaling 40% of Nvidia’s GPU sales
But who made up the other 60%? Here are three best guesses on Nvidia’s biggest mystery customer.
Tesla (TSLA)
It’s no small secret that Elon Musk and his ventures in Tesla (NASDAQ:TSLA) have an appetite for Nvidia GPUs. Musk openly announced on X that he expects Tesla’s purchases of Nvidia technology to total between $3 billion and $4 billion in 2024. That’s a sizable chunk! And it doesn’t include the prior expenses on GPUs that Tesla had made to send resources to Musk’s other projects at xAI.
With Tesla’s mission of making its entire car fleet fully self-driving, its consumption of Nvidia GPUs will likely continue. And possibly, it could increase depending on the efficacy of its data processing ventures. One example is Tesla’s Dojo Supercomputer. The brainchild is a cluster of 5,760 Nvidia A100 GPUs, giving a computing performance of 81.6 petaflops.
Through the Dojo system, Musk intends to route all of the constantly accruing and compounding data of Tesla vehicles on the road. This underscores the possibility that Musk’s combined AI ventures could make him Nvidia’s biggest mystery customer.
Dell Technologies (DELL)
No clear numbers exist on the amount that Dell Technologies (NYSE:DELL) has poured into Nvidia in recent years. But its latest partnership with the GPU giant could be a sign of extensive investment. This new venture openly drew attention back in May. DELL announced it would be using Nvidia’s Blackwell Tensor Core GPUs to expand its AI-Factory products.
These GPUs will cost between $30,000 and $40,000 and are likely the start of a new packaging partnership. Dell Technologies will integrate Nvidia’s top-of-the-line products into its data center computers. For example, the new Dell PowerEdge XE9680L uses eight Blackwell GPUs to provide the highest possible rack-scale density per device. This means exceptionally high linear processing speeds for customers relying on DELL’s data center platforms.
Also, the company will be announcing rack-scale solutions. Variants can support 64 GPUs in a single rack, with up to 72 NVIDIA Blackwell GPUs in a single rack for liquid-cooled variants. To make just one of these computing racks will cost DELL $1.92 million in GPU expenses, boding well for Nvidia.
Super Micro Computer (SMCI)
Super Micro Computer (NASDAQ:SMCI) directly relies on packing Nvidia GPUs into its data center computing products to stay competitive. In many ways, SMCI is a Nvidia reseller, but with the necessary supporting hardware and software for downstream customers to use Nvidia computers.
Consider SMCI’s recent announcement of its SuperCluster portfolio. These data center computing stacks can utilize anywhere from 256 to 512 Nvidia H100 or H200 GPUs, depending on cooling configuration and customer specification. For reference, a liquid-cooled SMCI SuperCluster using 512 H200 GPUs, which can cost up to $40,000 individually. And that would cost north of $20 million in just GPU expenses! Depending on the demand for this kind of computing power, SMCI could easily become one of NVDA’s biggest customers in the next decade.
Therefore, investors bullish on Nvidia’s revenue prospects can look to SMCI to understand the incredible lengths the industry is willing to go to integrate GPUs into its computing structures.
On the date of publication, Viktor Zarev did not have (either directly or indirectly) any positions in the securities mentioned in this article. The opinions expressed in this article are those of the writer, subject to the InvestorPlace.com Publishing Guidelines.
