Mr. Ives, thank you for joining us. So, Mr. Ives, AI party that started at 9:00 PM. So, I'd like to ask a few things about first, what time do you think it is now? Second, what key indicators are you using to judge the passage of time? Lastly, what signal in your view would mark the end of the party?
Yeah, great to be on with you as always. Look, I think it's about 11:30 p.m. Like I said, party goes to 4:00 a.m. And the reason I think that this AI revolution is still playing out, it's enterprises around the world. It's less than 5% have truly gone down the AI path and everything we're seeing in terms of data centers built, ship demand, I continue to think it's very early. Now, what would change our view is if we start to see projects shortened, we start to see demand soften. But to some extent, it's been the opposite. We've seen acceleration. So 11:30, we have a lot of time until 4:30 a.m.
Despite that, more than 80% of the S&P 500 companies have reported earnings. You pointed out that the real challenge to this is actually in bonds and the debt market. Can policy fix this?
Look, I don't know necessarily that it's going to fix it, but I do think that look, tech companies, let's call it 20 or 30% of the build out is going to be taking on some form of debt. So, in terms of yield, really that's definitely a huge risk, but I believe you're going to see yield ultimately come down. I think that's making the right decisions because also Bessent knows that if yields go much higher or considerably higher, that'll put a dent in growth for US tech companies especially relative to the buildout. I think that continues to sort of be a tug-of-war that's playing out in front of us.
Okay, let's move on another question. Nvidia is investing in AI customers and those customers then turn around and do business with Nvidia again, creating a circular financing risk. How do you think about the circular financing?
Look, I think some will say it's a risk. I view it as Nvidia is essentially building out the ecosystem of the AI revolution. So I don't view it as circular financing. I actually view it as this is just the start of really building out a new economy that they're really building out. Now, the circular financing argument really that's going to continue to come up. But in my opinion, that's something that I think investors have continued to be way too worried about it rather than the reality.
Nvidia has announced partnerships with major Wall Street firms to build a 500 billion AI infrastructure financing platform. Do you see this as an event that helps those circular financing concerns?
I think it does. Look, I think they've gotten validation from banks and I think to some extent any of the concerns that there's going to be a game of musical chairs and someone's going to be left without a chair, I think gets put to rest. I think the backing of all these banks, I believe everything Nvidia is doing is essentially building out the foundation of the AI revolution. And even though circular financing concerns will continue to be there, I think they've actually alleviated it again and again, not just with earnings, but in terms of the investments that they're making.
Compared to what you said about Nvidia, Nvidia is making an ecosystem of equity, but like if you look back, Nvidia stock slowed to two-digit growth this year, which is a clear slowdown from previous like triple-digit gains. And even Broadcom's strong earnings doesn't seem to be enough anymore in the market. How strong do you think these numbers should be in order to satisfy the market today?
Look, I think I mean looking at Nvidia's last quarter, right? I think investors went in there thinking it was going to be a shrug the shoulders. Yet those were numbers that I think were almost jaw-dropping, right? I think when you look at some of the results, Broadcom or some others in terms of the reactions day after the quarter or a few days after, I think it's shortsighted. I think you have to ultimately look at what this is saying about the broader demand when you piece the puzzles together from hyperscaler to chips to memory. And I think essentially what you're seeing is like demand's accelerating. And I think that to me is really the biggest sort of takeaway that we're seeing play out across tech.
Let's scope on AI demand. Recently, I really enjoyed listening to the Bloomberg podcast you appeared on. At the time, you assessed that semiconductor demand was 13 or 14 times larger than supply, but I'm curious about the details of how you arrived at that figure.
A lot of it is just the timing of chips. And if you look at the orders that are being put in, orders put in today, a lot of times you might not be getting those chips for another 12, 14, 16 months. It continues to elongate. I think when you triangulate it, that might even be underestimating what demand truly is. You know, in our opinion, this is all sort of pieces of a puzzle that are playing out. And I think when you look at memory, you look what's happening in Nvidia, you look what's happening Palantir, Snowflake, software companies and others, I think it tells you just a very robust growth story that's accelerated.
On the supply side, could the growth of China's semiconductor industry become a variable that affects your forecast?
Look, I think there's always concern about narrowing the gap in terms of China's semiconductor industry. Look, China's going to continue to be extremely competitive and what Huawei and all the chip players are doing. I continue to think that Nvidia and overall US chip demand is a few years ahead of where China is. But it also speaks to why Nvidia needs to be able to sell into China because by not being able to sell into China, what it essentially does, it actually makes China that much stronger, you know, in terms of their tech industry. I think that's what Jensen talks about. You know, this is an arms race that's playing out between the US and China. But for US tech companies, people want to come to let's say Apple and there was thought that maybe they would look to try to get some of their chips from China. That's going to be very difficult in this political climate.
Thank you. Also looking at the expansion of long-term agreements as well known as LTAs. Some argue that memory semiconductors are structurally breaking out of the typical cycle. Do you agree with that view?
I do. I mean in my opinion memory is almost going through what I view as a super cycle. Like so I don't think the typical cyclicality that you're seeing in memory I think essentially gets thrown out the window. We don't think that you have core equilibrium on the memory side probably till 2028, 2029. And I think that's what now look if you look like where memory prices, you look where the stocks are trading, there continues to be concern in terms of that dynamic but I don't think you probably don't have true equilibrium for probably about two more years.
