AMD, AVGO, AMAT & INTC: Stephen Sopko's AI Leadership Picks Beyond NVDA

SPEAKER_1 (0:00)

So now here to help us break down really what we're seeing from these developments, we're joined by Stephen Sopko, Practice Lead of Semiconductor and Deep Tech at HyperFrame Research. And so Stephen, always really appreciate your thoughts here. And so as far as when you assess the latest developments we've heard, I mean, broadly speaking, what has been your biggest takeaway and are you thinking that the innovation is keeping pace with all of the very high expectations for some of these names?

Stephen Sopko (0:28)

That's a great point. I think when Applied announced their Epic Center, this new way of doing research and development around chips back in 2023, 2024, it raised a lot of eyebrows. People were trying to figure out, okay, do we really need a new way of researching the materials that go into chips? And now fast forward to a couple of years later, they did their big unveiling at Semicon West just earlier this week. And their goal is to collapse the time to research the materials and the machines to make new chips from the 10 to 15 years it currently takes down to 6 to 8 years. And I think the thing we've seen over and over again is AI is now moving on an annual cycle. And it doesn't take much for it to outpace the ability of the industry to keep up and grow the technology along with it. So I think that's one of the huge things that came out. And a really good underlining point was another company in the space, ASML, releasing during their earnings talking about being capacity constrained. All of these companies, whether it's Applied, ASML, Lam Research, they're going to struggle to keep up with the demand. And one of the things that Applied said over and over again, and they had about $14 trillion of market cap on stage with them at their event earlier this week, Nvidia, Broadcom, Intel, you name it. Everybody was there, Micron, SK Hynex.

Everybody is talking about that ability to not only go faster, but make the processes more efficient. We call it yield increase so that you don't have to throw away as many chips. Because even when something goes absolutely perfectly, you have 500 to 1,000 process steps to make a modern chip.

And a lot can go wrong during those processes. So even when everything goes well, you end up throwing away a certain number of them. You reduce that number, we get more AI chips. So I think that's one of the biggest things I've seen come out this week, is that change in the way R&D and yield management is happening.

SPEAKER_3 (2:23)

Yeah, that's fascinating stuff. Stephen, just listening to you, it's such a breath of fresh air for us. Because look, we do this every day as you do, but this is such a complex space. And as you get further down and try to specialize and figure out what these companies do, it gets more and more confusing and difficult. Well, let's kind of do our best, I guess, to kind of go line for line with some of these stocks that of course our viewers and investors are interested in. One is Broadcom. Now, this was one of our kind of early winners. Nvidia was sort of that first mover, but Broadcom wasn't a ton behind it, but it seemed like it's sort of lost now relative. Look, I know the numbers are fine, but they're not as strong in terms of being able to raise their guidance the way that some of these other names are.

What is, I guess, standing out with Broadcom right now that's different today than maybe it was a couple of years ago?

Stephen Sopko (3:19)

Well, it's interesting. I think Broadcom is doing so much for so many different parts of the industry. You look at it, they're making custom chips for the hyperscalers. They're announcing a lot of products this week around the chips that allow systems in every kind of data center to talk to each other more efficiently using light. There's so much that that company does, that I think it's hard sometimes for people to really nail down which part of the company is acting at any given moment. The thing is the guidance from the company is phenomenal. I mean, we're talking these days about the overall semiconductor industry getting twice, 3X the size it is today in the next few years. I think that the key there is when you're a company like Broadcom and you are working with a ton of different partners, a massive supply chain, and designing the next generations of chips.

That is a story that needs to be told, and I think people just need to be able to separate. Okay, now we're talking about the part of Broadcom that works for the hyperscalers designing custom ASICs. Now we're talking about the part of Broadcom that is doing the photonics that let systems talk to each other and take it all as a combination.