Last week, electric-vehicle pioneer Tesla (TSLA 0.66%) turned in a second-quarter report that disappointed investors, who drove shares down nearly 10% on the following day. Revenue and earnings beat Wall Street's estimates, but investors were not pleased about the continued slide in Tesla's key profitability margins due to its ongoing vehicle price cuts.
As usual, CEO Elon Musk and the top execs who joined him on the earnings call provided some great information for Tesla investors. But Musk also made comments that were relevant to investors and potential investors in graphics chip specialist Nvidia (NVDA 3.46%), which is the main focus of this article.
Image source: Getty Images.
"Quasi-infinite" need for artificial intelligence (AI) training
From Musk's remarks:
[O]ur Dojo training computer is designed to significantly reduce the cost of neural net training. ... It's somewhat optimized for the kind of training that we need, which is video training. So, you know, we just see that the need for neural net training -- talking about quasi-infinite things -- is just enormous.
Musk uses the term "neural net training." A layperson could think of this term as "AI training" or "deep-learning training," as deep learning is the type of AI that's been experiencing rapid growth. A still simple, but more exact definition comes complements of Amazon's AWS site: "A neural network is a method in artificial intelligence that teaches computers to process data in a way that is inspired by the human brain."
The key thing to know is that Nvidia's graphics processing unit (GPU)-based products dominate the AI training market and the overall AI computing market. (AI computing includes two steps -- training and inferencing, with inferencing referring to the deployment of whatever is AI-trained...such as a car or a chatbot.) A June Reuters article estimated Nvidia's market share of the AI computing market at 80% to 95%, citing Wall Street analysts as sources. So, a "quasi-infinite" or "enormous" need for AI training is extremely bullish for Nvidia.
As for Dojo, Tesla is developing this supercomputer, which will use its custom D1 chips, to train its AI models to improve the self-driving capabilities of its advanced driver-assistance system for its vehicles. Tesla's top reasons for developing Dojo are likely to lessen its dependency on Nvidia and to help drive down its AI training costs.
If all goes extremely well in developing Dojo, Tesla could also sell these supercomputers or license the underlying technology to other companies. But that shouldn't scare Nvidia investors. The market for AI training will be massive -- large enough to support a few leaders. Moreover, Nvidia is so far ahead of the competition in this space that it seems poised to reign supreme for quite some time. It's lead is not only due to its hardware, or GPUs, but also to the software tools it has long provided to AI researchers.
"We'll actually take the hardware as fast as Nvidia can deliver it to us"
From Musk's remarks:
[W]e're also using a lot of Nvidia hardware. And we continue to, you know. We'll actually take the Nvidia hardware as fast as Nvidia will deliver it to us. Tremendous, tremendous respect for Jensen [Huang, CEO] and Nvidia. They've done an incredible job. And frankly, I don't know if they could deliver us enough GPUs. ... [T]hey've got so many customers. They've been kind enough to, you know, nonetheless prioritize some of our GPU orders.
The message in this quote seems to be that Nvidia has more demand for its GPU-based AI computing products that it can keep up with.
Indeed, Nvidia's recent fiscal first-quarter 2024 report indicates that demand for the company's data center family of products that enable AI applications is "surging," to use the term that Huang used in his earnings release statement. Data center revenue increased 18% from the prior quarter. Moreover, citing the strong demand for its data center products, Nvidia's management guided for fiscal Q2 revenue to soar 64% year over year and adjusted earnings per share to rocket 286%.
