Nvidia's demand is broadening and still exceeds supply, says BofA Securities' Vivek Arya
Joining us now, Vivek Arya, B of A security senior semiconductor analyst. And Vivek, it's great to have you here. It feels as if the enthusiasm behind the NVIDIA story, the multi years of of demand exceeding supply were confirmed for the most part. What are the headlines from the report from your perspective? Yeah, good morning. Thank you for having me. I think there is a global race that is going on between cloud service providers, between consumer Internet companies, between on premise enterprise, between the thousands of venture backed start-ups to take the best advantage of this wonderful and amazing new technology called generative AI or large language models. And NVIDIA is best positioned to deliver a turnkey solution to every one of these constituencies who wants to take advantage of this technology either to lower their cost or to improve their revenue profile. So what did we learn last night? I think 3 important things. First, demand is broadening. It's not just the cloud service providers. They went to only mid 40s percent of sales versus mid 50s the last quarter. So demand is broadening to automotive companies, to enterprises, to consumer Internet companies. Second thing we learned is that demand still exceeds supply. And I think this was a concern, Mike, going into the call because the companies in front of this major product transition to something called Blackwell. So there was this fear that there would be this air gap, but what we are learning is that just because of the breadth of product and the breadth of customers that they are able to go through this transition very smoothly. And the 3rd and very important thing, which I think sometimes gets understated is that the top line is not only growing over 200% a year, every dollar of sales is being converted to over $0.50 of free cash flow. Which mega cap company in technology or the market is able to do that while trading at only 1 1/2 to 1.6 times the multiple of the S&P 500. So I think it's it's checking every box of growth, you know, execution and cash flow generation. And the stock is still only about 30 times next year, multiple for what is basically a market weighted stock among most of the institutional shareholders. Yeah, it's it is remarkable. I mean, you say you see a a path toward perhaps $50 per share in in earnings power within a couple of years. So in the third year from here, I guess the the question big picture to me might be, you know, how long can a company sustain such a competitive advantage that they are given such a huge generous portion of the returns for this entire investment cycle at these margins and producing that much free cash flow? It just seems as if you know at some point you know capitalism goes after those margins. No, absolutely. But one thing we have learned is in semiconductors there is tremendous value to scale and incumbency and obviously you know just in technology there is tremendous value to being the 1st in the market with the leadership position. So if I look at the global accelerator market this year we think it double S to about 100 billion and Nvidia's 80% market share of that. We think it double S again over the next several years. And NVIDIA we think and maintain that kind of market share. And by the way this is not too uncommon. We saw similar multi year product cycles when the telecom service providers move from 2G to 3G, when they move from 3G to 4G, from 4G to 5G, there is always this three to four year upfront deployment cycle and I think they are just in the middle of that upfront deployment cycle. And the final point I would make there is that the price of entry into generative AI is not cheap. Just to have a useful deployment, you need to have at least a one or two billion of upfront CapEx because you need to have at least 10 to 20,000 GPUs to bring about a cluster to do anything useful. That is why these numbers are getting very big and NVIDIA is not just selling chips. They are able to upsell to boards, systems and software and that's why you see revenues get so big and we still think that they can sustain this for the next several years.