Why Nvidia doesn’t have all of ‘its eggs in the OpenAI basket’

00:00 Speaker A

Some viewers might be wondering, why does Nvidia’s Jensen Wong, why does he want to invest in Open AI and Anthropic in the first place? Like what’s the move? What’s the strategy?

00:13 Austin

Yeah, so that is a great question. So, the the way that I interpret what Jensen has said all along is that Jensen believes that the his vision is that the whole world, every industry is going to be transformed by accelerated compute by AI. And so if you put yourself in his shoes, your question is, how can you ensure that that vision happens? Obviously, these AI model companies are a big piece in creating that future. They obviously have to train the models, inference them, get them into the hands of customers, support them, you know, build consumer products around them. And Jensen is as an infrastructure platform, you know, he’s really incentivized to help make sure his partners can bring value to end customers. Um so, of course, Nvidia has lots of money to invest and so that not only are they investing and making their platforms better, but it would make sense that they would invest in their partners to ensure that their partners can, you know, bring that future to bear as quickly as possible.

01:31 Speaker A

If Open A if OpenAI stumbled, Austin, and who knows? I mean, Google Gemini, you know, coming in hot. If Sam Altman stumbled, how exposed would I be to that stumble as an Nvidia investor?

01:52 Austin

Yeah, that is really an interesting question. So, you know, what Nvidia investors should feel confident about is that over the past several years, it we’ve seen diversification. It’s not just Open AI anymore that makes these models. It’s obviously Open AI, it’s Anthropic, it’s Google, of course, but then there’s also open source uh options. and then even XAI for example, you know, came into the race late and is able to kind of catch up and participate. Obviously, Open AI is a very big customer. They’re driving a lot of usage. Microsoft buys a lot of Nvidia GPUs and it’s to serve Open AI’s models. So if Open AI were to stumble, that would obviously impact demand for compute across the industry. On the other hand, they aren’t the only ones. Google’s out there, you know, Google of course has their own TPUs. They also uh buy Nvidia GPUs and rent those out as well. And so Nvidia doesn’t have all its eggs in the Open AI basket, but of course, there’s definitely quite a pile of eggs in Open AI’s basket.