Dive into the world of semiconductors and Automatic Test Equipment with our educational podcast, Advantest Talks Semi, where we explore the power of knowledge in this dynamic field. Hosted by Keith Schaub, Vice President of Technology and Strategy and Don Ong, Director and Head of Innovation for Advantest Field Service Business Group, at Advantest, this series features insightful conversations with experts and thought leaders in the industry.
In today's fast-paced environment, continuous learning is essential for staying ahead. Join us in these thought-provoking discussions, where you can learn about the latest trends and cutting-edge strategies being used in the semiconductor industry. Explore how innovative technologies are revolutionizing testing processes and shaping the future.
Stay updated on the ever-evolving semiconductor industry with Advantest Talks Semi, and gain exclusive insights into the future of technology.
The views, information, or opinions expressed during the Advantest Talks Semi series are solely those of the individuals interviewed and do not necessarily represent those of Advantest.
From Silicon to Systems: Redefining Competitive Advantage-Part 3
•Season 3•Episode 8
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
0:00
|
22:11
In today's episode, Advantest's Dong Ong sits down with Salah Nasri, CEO & Co-Founder of ISIG (International Semiconductor Industry Group), to discuss how the semiconductor industry is redefining competitive advantage. @SemiconductorLeadershipPodcast
Subscribe to Advantest Talks Semi Podcast for the latest episodes and stay ahead of the curve with cutting-edge insights from industry leaders.
If you enjoyed this episode, we'd love to hear from you! Please take a moment to leave a rating on Apple Podcast. Your feedback helps us improve and reach new listeners.
Don't forget to subscribe and share with your friends. We appreciate your support!
So, Salah, for decades, the semiconductor industry focused heavily on transistor scaling. Looking ahead, where do you think the next major sources of competitive advantages will emerge?
SPEAKER_01
A few areas stand out. And what's interesting is how different they are from each other, right? It's not a single frontier, without a doubt. It's uh several converging ones. Um, of course, I'm gonna have to be very biased, and uh my our business is successful because of advanced packaging. So I'm gonna start off with advanced packaging, it's probably the most immediate. Um, the ability to integrate heterogeneous chiplets into a coherent system is becoming a genuine differentiator, and it's a space where equipment, materials, and the process know-how all intersect in interesting ways. Um there was a I can't remember where the exact report was from, but um it basically projected that the advanced packaging market um is without a doubt going to surpass that $100 billion, right? And this is by 2030 onwards, and that is eightfold if we compare it today. So, and this is driven specifically by AI demand. Now, what we need to take in consideration, Don, is that this is a structural shift in where value is being created in the stack, not just the incremental growth. So if we look at AI-driven design and verification in another area, I would highlight that the companies that can compress development cycles through intelligent automation will have a real edge, not just uh uh on time to market, but on the quality of what they can actually build. And increasingly, I think the software and data layer is where sustained competitive advantage gets locked in as well. Um, period. I mean, look, the the chip is the foundation, but the platform built on top of it, if we talk about the tools, the models, the developer ecosystem, is where the uh uh I'm trying to think of the right word, the stickiness accumulates, right?
And the other key one is is co-package optics. I mean, you know, if you look at the last six months or so, the key players in the market are snapping up these silicon photonics companies like know tomorrow, right? You have the you know Marvell acquiring Celestial AI not that long ago. Um, the uh Crado buying um I think it was Dust Photonics the other day, uh Teramount being bought someone by someone else. I can't remember the exact name. You know, uh co-package optics is also gonna be a key one, without a doubt. You know, um this is um the solution to the power problem that's being faced, and also the data, you know, along once upon a time, you know, we all depended on you know electrons being the flow of data. Now it's light, right? Photons transferring all this data, right? So this is something as well that's gonna be the next major source. And if um if you wanna double well, if you want to double check my um statements, just have a look at the photonic stocks today and get back to me. Yeah, have a look at companies like AXT, who's supplying the material for Indian phosphide. You know, have a look at lamentum, have a look at coherent, right? You know, have a look at all of these uh players in the photonics world, just look at the stock market and how they're doing with their pricing because they're being priced in for what they're gonna be able to do
in the future. And you know, when people ask me about you know, how do you analyze the stock market? The best analogy I can have or say is that if you have a glass, right? And Don, if I said to you, give me a beer, a beer can, and I poured the beer into the glass, what happens is that you have the liquid at the fraction part of the glass, and the remainder part will be foam. Now, in the stock market, the best way to explain it is that beer liquid part is the actual revenue of the company, the value of the company, etc. And the foam is the hype expected, expected, right? It's the hype expected, it's the potential that it can achieve, and etc. And normally the foam is usually much bigger than where the actual value of the beer is, but that is the expectations of the emergence of photonics. And if we look at co-package optics, this is going to be the next um key competitive advantage that the industry will
So you mentioned a lot about advanced packaging. So I want to unpack that a little bit. I know it's it is the hot topic right now. Everyone's talking about advanced packaging. In your words and in your in your view, your perspective, advanced packaging is very often right now described as strategic rather than operational. Why has packaging become so important in the AI era? I know a lot of people have different definitions I want to hear your definition.
SPEAKER_01
My chairman is gonna choke me because he's a top guy in the advanced packaging side of things. So I need to make sure that I get this right. Well, look, advanced um advanced packaging is essentially redistributing the value that used to sit entirely at the leading edge node, right? Once upon that, everyone's front-end, front end, front end. Back end was like the disgusting part. You know, you don't, it was the dirty side, you don't talk about the the the back end, right? Now it's more just as important more than ever. And you can now achieve meaningful performance gains through integration and interconnect innovation that previously required a full node shrink done. So that changes the competitive landscape significantly because it gives more players a credible path to differentiation. So you know to put some scale around it, the packaging costs for uh top-end AI chips could climb to nearly $1,300 per chip in a couple of years before falling sharply. Uh, as the assumption that we know is panel level packaging matures, and the cross trajectory tells you exactly where the complexity and the value is migrating in the stack. So what's happening is that AI is having a different but equally structural effect, it's creating enormous demand concentration around a small number of compute architectures, which gives the customer, which is the hyperscalers, unprecedented influence over roadmaps. But at the same time, it's accelerating the need for custom silicon. So you know, you're pulling more companies into chip design who weren't there before. And you have to you have both consolidation and fragmentation happening simultaneously, which makes for a very fascinating competitive dynamic. And as you know, I I I'm so lucky, Don. I get the fortune to rub shoulders with the guys behind the creation and the development of EMIB and Foveros and Coos L and Coos R, right? The guys that were behind it have now just recently retired, and they say to me, Salah, this is currently now, now and the next years to come, is the most fascinating, most incredible time to be in advanced packaging. And they're still involved, even though they're retired, they're still involved in the game. So that just says a lot about where it is today.
SPEAKER_00
Yeah, and and you alluded to this, right? You mentioned this at the very beginning, is that in the historically, and for this industry, the value itself is more concentrated in the front end, and now with packaging, it is the value is now being redistributed a little bit, right? And we see now it move towards integration, packaging, software, and platform. So
for the for this data and software is becoming so important. Where how do you see software reshaping semiconductor value chain over the next decade?
SPEAKER_01
Well, yeah, you know, the again we have to pick the elephant in the room, right? Agentic AI. You know, agentic AI is uh is is changing the world. You know, uh just recently I've just started using Claude and it's mind-blowing when it's done. You know, what it does for me on a productivity standpoint of view. Um, I think um, you know, with the you know, you have these elements of digital twins and etc. But I think the agentic AI point of things, from a software standpoint of view, it's gonna be uh it's already shaken up. You know, imagine once upon a time, Don, right, to create an app would take you months to make, right? Now with agentic AI, it can do it for you in an hour. You know, you could build a website in half an hour. You can it's remarkable what agentic AI is doing today, and I think from the software standpoint of view today, you know, already it's as complex as it is, but with the addition of agentic AI, the impossible is gonna be possible.
SPEAKER_00
I like that. I like that answer. So I'm gonna ask, I know you talk a little bit about this, but I'm gonna re- I'm gonna ask this question from a different
angle. So we talk about silicon photonics, advanced packaging, co-package, heterogeneous integration, and and agentic AI that is coming in, and we're gonna change our industry, gonna bring a lot of value to our industry. But what technologies do you believe will fundamentally reshape the industry? And also which might be overhype?
SPEAKER_01
This is a tough question. Um there's uh a professor, a very dear professor of mine, Baba Shiv, Professor Baba Shiv at Stanford. Um he said to me the person that can fix the energy problem will be the next trillionaire. I think this is the biggest one. And in terms of hype, um I need to be careful here, Don. You're gonna get me in trouble. Uh in terms of hype, I don't you know what, to be frank with you. I mean, you know, I just got my hands dirty in the quantum world, in quantum computing. Um there's a lot of hype around quantum, right? Um but for quantum computing to succeed, there needs to be a convergence with the semiconductor industry, period. You know, um the capability of using you know CMOS chips in the world of quantum um says a lot, right? Um until there's an actual application use, I would say the quantum side of things are hype, right? Um, but everything else that you mentioned, I don't believe it's hype. Yeah, code package optics is legit, silicon photonics is legit, advanced packaging, heterogeneous integration is legit. Cool. I like that. And so I'm so I'm swaying away from your question. I know what you're trying to do here, but uh I'll try to that's how we make it interesting. I like these controversial I like these controversial questions, but um nah no no and no no chart. I mean, you know what? I mean, look to be frank with you, not many people would would notice. You know, we struggled when we first started in the business. You know, we're a 16-year-old association. The first five years, you know, we were jumping in logic and memory and all of this type of stuff from an agenda perspective that we do for our summits globally. And you know, it's like 2015, you know, there was no other association that could get TSMC and Intel in one room, but we were capable of doing it. And because it was on advanced packaging and we're focusing mainly on that, there was much more ease of sharing information because it it's a it's a secretive game, the industry. Your customers, advanced customers, are very secretive in the work that they do. They don't want to show out their secrets, they don't want to, you know, elaborate. But you know, the advanced packaging perspective created an open-minded or open room discussion um to help it get it where it is today, right? Um, you know, nobody would have imagined, you know, I'll give you a good example. You know, I had a meeting in January, actually, uh, January, mid-January with uh CC Way, the CEO of TSMC. And um even he was saying to me in specific, he said to me, you know, Salah, the we put a lot of money in advanced packaging, right? And I was struggling to see the ROI. You know, and then all of a sudden, the COVID situation and AI comes in, they're doing billions per quarter in advanced packaging, right? Um, so advanced packaging, heterogeneous integration, co-package optics, this is not a hype.
in that. Moving towards the end of episode three, I have to ask this question. You you have very privileged perspective, and I'd like to see this. If you were to look 10 years ahead, what do you think the future semiconductor leaders will need to understand that traditional semiconductor leadership models may have overlook?
SPEAKER_01
Good questions, man. Uh things have changed. Things, times have changed. And I think in 10 years, um, that mindset shift. I mean, you know, the the guys that are retiring now are the ones that were here 30, 40 years ago, right? So the new set of leaders are guys like us that we don't have much gray hair. We have a few grey, I mean I have a few gray hairs in my beard, and you know, like like you know, with the the mid-age folks, we're gonna become the front of the game, right? Um, and I think there are three things. There are three things, and I'd argue all three all three are equally important. Um, and you can't get two out of three and expect it to be enough. First, ecosystem clarity. What does this mean? It's understanding precisely where you create unique value and building that trust, as I explained early on, those that partnership to complement it rather than trying to own everything. So focus is key. And this sounds obvious, but um a lot of companies, Don, a lot of companies still haven't fully made that shift, and the instinct, of course, is to expand vertically is strong, and it has to be consciously checked against a clear-eyed view of where your irreplaceable value sits at the end of the day. The second perspective will be adaptability and capital allocation. So what we see today, Don, is that the technology landscape is shifting fast enough that rigid uh 10-year roadmaps are a liability. And any industry leader today needs to strategic need they need to have strategic strategic flexibility built into how you invest and the ability to redirect when all of a sudden some sort of situation changes rather than defending sunk costs, period. Um back to the poker game. You're gonna know when to fold. Alright? Um and third, and this one I find most uh underappreciated, um, is talent and knowledge continuity, which is why, you know, uh we you know, I I thank you and and and the guys at Advances for doing a podcast
like this. I have my own podcast above me, you can see it, semiconductor leadership podcast. And this is um creating a way to create generosity and knowledge sharing. You know, if you look at it with you know with fewer than 50,000 engineers globally who have the specialized knowledge to manufacture leading edge chips, the depth of human expertise in an organization is a genuine competitive asset. This is what we call IP done. And the companies that know the companies for me that know how to retain, how to develop and attract that expertise over the next decade will have an advantage that is very hard to replicate quickly. And I think more than ever today, more than ever today, Don. You know, once upon a time before it was very, you know, we were a small clique, the semiconductor community. Now, you know, I think it's more important than ever that any semiconductor industry leader needs to have a personal brand. You know, they need to have a strong LinkedIn profile, for example, right? You know, they need to have strong presence out there, you know, they need to be involved in podcasts like this, for example. Because at the end of the day, the next generation, with all due respect to exhibitions or large-scale conferences out there, the next generation are listening to podcasts. This is where they're gonna pick up the that's their form of education today. Yep. Yeah, they're the even the yeah, even the I mean, you know, I I wrote the book because of uh uh uh it was a personal challenge on my side because I had it frustrated. My frustration was that I couldn't afford as a business, I couldn't find afford veterans in the semiconductor world to come join the organization that I run. So what I was doing is basically I'll find you know people that just come out of university, for example, or um people that are in other industries and they're frustrated where they are, right? But when they join the organization, I have to explain to them what the semiconductor ecosystem is about, how it starts, when did it start, why is it important, who are the key players, and etc. And that was the reason for me to write that book because whenever someone came in for an interview, the first thing I say to them is before you have my an interview with me, you read the book and then you come talk to me. Because by the time they read the book, they'll come to me and they'll know the industry very, very well. And it helps me with the training process in terms of education, right? So I think these are the the from a personal brand point of view, I think it's more important than ever that um industry leaders also need to take uh into consideration methodologies that helps out with talent and knowledge continuity. Very underappreciated, Don. So um, I hope that answers your question because at the end of the day, you know, the future of the semiconductor industry will be defined by those who can integrate strategy, who can integrate technology and ecosystems into coherent systems of value creation. That's what it's about. Value creation.
SPEAKER_00
And those are the key elements that get you there. Yes, yes. I 100% agree with you. And and that's also our in in the in a cliche is also that if our industry is moving from a chip-centric into a system-centric, and that system is not just on the chip itself, it includes the people behind it. So I like that a lot.
With that, a huge thank you to Salah Nazri for joining us and sharing his perspectives on the evolving semiconductor landscape. And thank you all for listening to Avantes Talk Semi. Until next time, stay curious, stay innovative, and keep pushing the boundaries of what's possible in semiconductors. All right, Salah, thank you so much.
SPEAKER_01
Thank you, Don. It was a pleasure um being on this. And as I say on my own podcast, um, it takes a village to raise a chip, and also it takes a village to raise a podcast. So don't forget to hit that subscribe button uh for the Advanced Talks podcast. Thank you, thank you so much.
.jpg)
