Artificial intelligence (AI) gave Nvidia (NVDA -0.01%) stock a big boost in 2023 thanks to the terrific demand for the company's data center graphics cards, which are playing a central role in the training of AI models. And it looks like the windfall gains for the chipmaker aren't going to stop.
Electric vehicle (EV) maker Tesla (TSLA -8.78%) is likely going to give Nvidia's business a nice shot in the arm over the next year and contribute a substantial chunk of the company's data center revenue. Let's see why that may be the case.
Tesla could add at least $3 billion to Nvidia's annual revenue
Tesla recently pointed out that it will start building its AI-focused supercomputer, christened Dojo, this month. By January 2024, Tesla plans to equip Dojo with 100,000 of Nvidia's A100 data center GPUs (graphics processing units). The EV maker plans to eventually deploy a whopping 300,000 Nvidia A100 GPUs by October next year, giving Dojo a massive computing capacity of 100 exaflops.
That's huge considering that Tesla's current supercomputer, which is used to power the full self-driving (FSD) capability in its cars, was reportedly using nearly 6,000 A100 Nvidia GPUs in October last year, which gave it a computing performance of 1.8 exaflops. This large-scale ramp-up of Tesla's AI infrastructure is going to be a big boon for Nvidia.
The A100 data center GPU reportedly commands a hefty price tag of $10,000 to $15,000. Multiplying that number by the number of A100 GPUs that Tesla is intending to deploy indicates that Nvidia's top line could get a $3 billion to $4.5 billion boost over the next five quarters. Nvidia's data center business has generated just over $19 billion in revenue over the past five quarters, so Tesla alone could give this segment a 15% to 23% boost over the next five.
However, Nvidia's revenue opportunity from AI supercomputers isn't just limited to Tesla. A closer look at the chipmaker's prospects in this market will tell us that it could be scratching the surface of a lucrative long-term growth opportunity.
Supercomputers could drive healthy long-term growth for Nvidia
The AI market is expected to grow at an annual pace of 39% over the next decade, according to Precedence Research. As a result, the need for AI-focused supercomputers is likely to increase at a nice pace as well since they will be needed to train huge and complex models because of the large amounts of processing power, storage, and memory they are equipped with.
This explains why the high-performance computing (HPC) market, which includes computing solutions provided by supercomputers, is expected to generate annual revenue of almost $97 billion in 2028, compared to this year's estimate of $57 billion, according to Mordor Intelligence. Nvidia has already positioned itself to tap this massive opportunity by building a solid client base.
We have already seen that Tesla is going to be a substantial contributor to the semiconductor giant's revenue over the next year or so, but the Elon Musk-led company isn't the only one relying on Nvidia GPUs for its AI ambitions.
In January 2022, Meta Platforms introduced its AI Research SuperCluster supercomputer to train large and complex AI models. The supercomputer was powered by just over 6,000 Nvidia A100 GPUs at that time, though Meta added that SuperCluster would eventually be powered by 16,000 GPUs once it was completely built.
Earlier this year, it emerged that Nvidia's A100 GPUs have been deployed to train OpenAI and Microsoft's popular chatbot, ChatGPT. According to market research firm TrendForce, ChatGPT could end up deploying at least 30,000 A100 GPUs, which would translate into a nice chunk of revenue for Nvidia. Citi analysts, on the other hand, gave a more ambitious forecast earlier this year, estimating that ChatGPT's adoption could boost Nvidia's revenue by $3 billion to $11 billion.
All this indicates that Nvidia is the go-to provider of data center graphics cards to companies looking to train AI models. Not surprisingly, the company reportedly commands a whopping 95% of the market for machine learning GPUs as per New Street Research. So Nvidia should ideally enjoy solid growth in demand for its data center graphics cards for a long time to come, and that should drive healthy top- and bottom-line growth for the company given its strong pricing power in this space.
NVDA Revenue Estimates for Current Fiscal Year data by YCharts
Nvidia is trading at 55 times forward earnings, which is expensive compared to the Nasdaq-100's forward earnings multiple of almost 30. However, growth investors might want to consider buying the stock even though it is trading at a rich multiple because of the potential growth on offer, which could help Nvidia sustain its hot stock market rally in the long run.
