Name: Packaging, Test, Market Adoption
Uploaded: 2025-10-17T17:00:00.000Z
Duration: 24 min 38 s

Clips
More from theCUBE + NYSE Wired: GSA Series - Women in Leadership

Cherry Ooi

Director of Sales

ASE

Nitza Basoco

Technology & Market Strategist

Teradyne

Pam Fulton

Sr. Principal Engineer

Intel Foundry

play_circle_outline Panel discussions featuring experts from ASC, Teradyne, and Intel.

play_circle_outline Overview of the semiconductor industry and its impact on technology and infrastructure.

play_circle_outline Unlocking the Future: AI, High-Performance Computing, and Advanced Packaging in Semiconductor Growth and Efficiency

play_circle_outline Yield and manufacturability as critical factors in semiconductor production.

play_circle_outline Future trends in packaging, like co-packaged optics, driving technology innovation.

Info
Transcript

Packaging, Test, Market Adoption

Cherry Ooi

Director of Sales ASE

Nitza Basoco

Technology & Market Strategist Teradyne

Pam Fulton

Sr. Principal Engineer Intel Foundry

In this GSA Series – Women in Leadership segment, theCUBE’s John Furrier sits down with Cherry Ooi (ASE), Nitza Basoco (Teradyne) and Pam Fulton (Intel) to unpack the realities of “panel #1: packaging, test and market adoption.” The discussion dives into how AI, high-performance computing and edge demands are driving advanced packaging – 2.5D/3D integration, chiplet designs and fan-out – and why OSAT/foundry collaboration is now a system-level exercise, not a point solution. The guests explain how faster device turnarounds and compressed cycles are reshaping ... Read more

explore Keep Exploring

What is the current demand trend for chips and who are the guests discussing this topic? add

What are the current market dynamics and factors influencing the semiconductor industry, particularly regarding computing performance and product release speed? add

What are the major areas of growth identified by ESC and what advanced packaging technologies do they provide to meet customer demands? add

What are the key factors influencing yield and manufacturability in advanced packaging and how are companies addressing these challenges? add

What changes are expected in the market from a packaging perspective that may not be obvious now but could become clear in the future? add

bolt Powered by CUBE AI

Packaging, Test, Market Adoption

>> Hello, I'm John Furrier, host of theCUBE here in our Palo Alto Studios for a special series with theCUBE and the NYSE Wired program and community. We're featuring the GSA, the Future of Semiconductors, the Women Leadership Series, looking at the focus of how semiconductors are changing the world. Of course, there's so much technical going on. We're going to feature multiple panels. And panel number one is the packaging test and market adoption. This is where we're going to see the performance of how the demand for chips on the performance side, as well as in the new systems has more demand than ever before. And we're going to talk about all the action there. We've got three great guests, Cherry Ooi, director of Sales at ASC. Thanks for coming on. Nitza Basoco. Sorry, I got that right. Tech and market strategy at Teradyne. Thanks for coming on. I had it right the first time. We'll get it right next time.

Nitza Basoco

>> No problem.

>> Pam Fulton, senior principal engineer at foundry at Intel. Congratulations on all the good news you guys are having coming out of Intel.

Pam Fulton

>> Thank you.

>> Thanks for coming on theCube. This is a super hot area. As you know, we've been covering the focus of semiconductors, have been changing the world. You can't read a headline these days without seeing news around large scale data centers. Billions and billions of dollars of spend. Some have been calling it the next build out of infrastructure similar to 4G and 5G. You're starting to see the AI headlines, a lot of hype, but the reality is kind of coming on, but makes it all work as the semiconductors. This is what you guys are doing. And the packaging is a key part of it. So I guess the first question, Cherry to you is the market dynamics is the key because what's changed? Because on, accelerations happened on the computing side. Also, the speed of product releases. These seems to be big things we talk about all the time. What's going on in the market?

Cherry Ooi

>> So from ESC actually, we do see that the three major growth that is coming is actually coming from the AI, from the high performance computer and also the edge computing. So these are the ones that actually require a lot of higher bandwidth, lower latency, and higher power efficiency. So here is where we actually comes in and as an ASC, we actually offer all this advanced packaging like 2.5D, 3D integration, packaging, fan-out technology. That is the one that is actually used to actually meet the customer demand on this very complicated chiplet integration design. So that is where we are.

>> Nitza, I'll get your perspective because we were talking before we came on camera. We're all like fans of the AI infrastructure, infrastructure in general. It's the hottest area right now and a lot of people don't know what's going on under the covers, even down to the lowest level. If you look at some of the software engineering techniques, you're seeing old school computer science, it now is going so closer to the hardware. This is a huge market dynamic, but also developers are getting more freedom with Co-pilots and code assistance. So you got a coding revolution happening, but also a chip revolution. What's your perspective on the market?

Nitza Basoco

>> So it's funny because you think about software defined devices and what's going on there and when you're looking and understanding the trade-offs that what you need to do at software, with what you need to do at hardware and how can we work together to develop a device that's not going to only function in a tester or on a board, but in the system in which it's going to be in. And I think that's really changing how people are looking at it and defining how devices are created.

>> And the convergence of test packaging is happening too. It's almost like compressed.

Nitza Basoco

>> Yeah. So if you think about it, so tests used to outpace Moore's law and we had a lot of things for site counts as well as the different types of techniques that you could put inside of it. But now, with process notes slowing down, you're thinking, okay, so tests should be okay. The thing is advanced packaging has taken over and because of that, that allows for faster turnarounds for the devices, which then puts tests in a very difficult position. So we have to actually do things at a lot faster pace than we're used to. And then there's new techniques coming out. We have new co-packaged optics at different types of techniques that are coming from the packaging technology, that we have to now go ahead and be able to do something about it.

>> We all marveled at the iPhone and how devices get faster and all the compute on there, it's only going to get better because now training and inference is going to come to devices, phones, wearables, cameras on, light poles, computer vision, this is what the market is pointing to. Right now is obviously going there fast. Pam, a lot of people don't understand how you make these things, like the chips. There's fabulous ... Explain kind of what's happening at the production site, because at Intel and certainly in the US there's a lot of discussion around how do we build these things? And even today in the news, I saw some news on Intel this week, there's a lot more going on in Arizona. There's so much action happening. What's the difference between the fabless and the fabs? Unpack that for us.

Pam Fulton

>> Sure, certainly. So in order to manufacture semiconductors, you need a few things. Number one, you need silicon. So silicon, there's a few really big foundries in the world. Number one of course is TSMC, similarly is Intel. And so we manufacture all of the silicon. There's a lot of integration that happens to the silicon level. All of the design for the computer architecture goes into the silicon itself. Advanced packaging comes in where you actually have to connect the silicon to the outside world, whether it's to a motherboard that goes into a PC or into a data center or something like that. So advanced packaging enables not just the silicon to talk to the outside world, but also to talk to memory and other components. And so all of the fabless companies do their silicon design and then they work with a foundry or an OSAT like ASC or Intel in order to design their packaging, something that will complement the silicon as well as their end use application.

>> The market seems very interesting right now seeing Jensen Huang on TV, Lisa Su from AMD. It almost seems like a team sport. Now, people might not know this, but semiconductor is kind of a small circle. There's a lot of teamwork involved. You guys have customers if the cycles are long, they're shortening.

Pam Fulton

>> Yes.

>> Talk about that dynamic because there's a lot going on between the relationships between companies in the industry. Anyone want to jump in on that and take that question?

Nitza Basoco

>> What's funny enough is that Pam and I met at the Chiplet Summit a couple of years back and one women in Chiplets but also because I'm in test and she's an Intel in process and was trying to figure out how do we partner together. And that's the best part about it. We partner with Intel. We partner with ASE, we partner with TSMC, we actually partner with the whole ecosystem. It's really important because people are getting all sorts of test challenges. They're getting packaging challenges and no one company can do it by themselves. So we have to come together and figure out how to do it to address those.

>> Yeah, and also supply chain, we're all dependent upon the same thing. Parts got to be available. I like the expression, "Everyone's along for the ride." It seems to be the quote I like the most. Talk about the customers that you guys have because I think again, look at the customers. They're going in planning cycles. They're trying to figure out that next device to be fast, the customer readiness, the decision making. Can anyone share the thoughts on the mindset and also just the environment of the customer. There's a lot more technology involved. We're seeing a lot more kind of back to the 90s kind of mentality. It's a famous quote, we get on theCUBE a lot where memory has to sit here and there's ... A lot of systems thinking going on. Anyone want to talk about the customer side?

Cherry Ooi

>> Yeah, actually, I can add on a little bit. Especially people talk about the chiplets, right? The chiplet is where actually you put everything, every chip, any function together in one single chip. So it's actually, it's a game changer, what we see. Right now, this chiplet is actually forcing the whole industry to shift from a product design to the system level solution. So at ASC we actually no longer just selling packaging or testing. We actually are co-architecting the system level solution with the customer. So this actually ... That's why we actually ... Good thing is that ASC has the foresight that we actually already developed this advanced packaging 10 years back. So that is what we see from the customer needs, but I guess is that the most important thing is to really work with the customer very closely to understand their needs.

>> And making those bets are critical.

Cherry Ooi

>> Yes.

>> And knowing what's going to come. It's a longer cycle. Nitza, talk about the trade-offs because I've had many conversations around. Well, if use this or that, this is what I get for, and then people are trying to figure out how to optimize. Talk about the complexity side of this, there's trade-offs for what you solve for.

Nitza Basoco

>> That's a rough question.

>> I didn't mean to put you in the spot.

Nitza Basoco

>> I know, but that's fine. Actually, a conversation we have with a lot of our customers because they're thinking, "Hey, I need to do it in a certain manner." It's like, "Well, okay, there's a certain flow for tests that happens." But then if you step back and do an analysis from a fiscal standpoint, so a cost of ... The standard cost of the device, then you can think of different types of trade-offs because it's not just the cost of tests, it's the cost of quality, it's the cost of yield and how does that affect me? Should I be inserting a test flow? Should I not be inserting a test flow? Should I be maybe flipping them around? So that makes more fiscal sense, but I still have the test coverage and the quality level that I need.

>> Pam, I want to go to you. She mentioned yield. Yield and manufacturability is a huge concept. We had a former Intel exec, Diane Bryant on multiple times and she always talks yields. She loves to talk about yields. Obviously, she bleeds Intel, she's a long-time veteran. Talk about the manufacturability because you have to make these. Yield matters if it fails, if you have more failures than positives, yield and manufacturability, can we do it? That is a whole other dimension. What's your thoughts on this?

Pam Fulton

>> It is, and part of the advanced packaging ecosystem that foundries and OSATs have been going through is really developing a lot of early yield indicators using things like test chips or test vehicles we call them, in order to look at IPs and look at packaging techniques and make sure using our test methodologies that we can actually get yield results before the product even comes to the manufacturing line. And then, we have continuous monitors all throughout the manufacturing process to try and get a good predictor of what it will look like. And then built into the silicon design itself, there are techniques that we can use actually in the field in order to be able to look for issues before they impact the end customer.

>> Cherry, talk about supply chain readiness because that's a big component because it has to be managed and it's super important. What's the challenge there? How do you guys look at that and what are some of the things you guys are helping customers meet the challenges because the demand for AI and high-performance supercomputing is at all-time high.

Cherry Ooi

>> Yeah. ASE has been actually on the forefront of supporting ... Enabling the AI and the high-performance computer. So there are so many elements. It is actually not easy to be on the forefront on this area. But on the scaling of manufacturing, I will narrow down into one key component, which is actually the capacity expansion or the capacity readiness. So oftentimes, as you know, advanced packaging require a lot of investment and actually you also require a lot of enormous of the production space. So we will have to actually work with the customer very closely in the beginning to really understand what their needs, two, three years down the road. So that we can actually have the lead time to prepare of having the manufacturing line to train the people, to bring in the equipment, that requires us to actually work closely with the customer. So that again, I think to support the AI and the high-performance computer that's moving so fast, having the flexibility and the foresight is very essential.

>> Nitza talk about the ... She mentioned AI. There's a phrase that we always use, AI for this and this for AI. So it's AI for networking and networking for AI. How are you seeing AI and machine learning being used to accelerate? Because the supply chain readiness question reminds me of the old COVID days, not to oversimplify, but you go buy a big Mercedes or a big car and you can't get it because that one ship didn't ship. Like, "Okay, we can relate to that," but this is similar problem. If you're on a cycle and that part doesn't work, I mean you're dead in the water.

Nitza Basoco

>> You're dead in the water. That's correct. And so everyone's thinking, "Oh, AI for semiconductors, AI for AI and AI for test." And a lot of our customers are coming to us and saying, what are you guys doing? And we're looking at it more holistically and saying, okay, what does Intel need? What does ASC need? What does our customers need? But also what does Teradyne need? And so you do it from different perspective. You look at the hardware. How do our customers get their hardware faster? Can AI help in that capacity? They have to create test programs. Can they get the software done faster? But more importantly, when I get a device, I get the debug done quickly and can I then connect it with maybe things I've seen in the past, not seen in the past? This is where AI can really help.

>> It's interesting, I always have this gut feeling that the way AI has been working is once you start using it, and I know you guys are using it, all you guys are. And it's like once you do it, then you discover new things and that yields breakthroughs.

Nitza Basoco

>> Yes.

>> Any observations that you'd share on what comes when you start using some of the tooling? Does it change some of the mechanisms? Does it change some of the process?

Nitza Basoco

>> Actually all of it. So the processes get changed because people are realizing they can do things in a different manner, in an innovative way in order to shorten down their time either to production or time to market. And then it's thinking, "Oh, if I could do this, maybe I could do something else." And so they need just starts snowballing really quickly because then you get excited that you're trying something new and then you go down a rabbit hole of figuring out, well, what other new cool things can I do? And then all of a sudden your process are more efficient, your type to market is faster.

>> Yeah, I mean I have a weird philosophy. I think that if the innovation comes in there, new ways to manufacture is why I was asking about the manufacturability Pam. Because if people discover new ways to do things that could change the future of tape out or how we work together. So Pam, I guess the question to you is the future tape out is one question, but they mentioned team sport. Data sharing and security is huge too because now the attack service area, if you can secure at the root level of the physical level, good things happen. So physical sharing and security and then the future of tape out as things change.

Pam Fulton

>> It's a really complex environment because something you mentioned earlier about collaboration, enabling some of these AI agents and things like that, you really need data sharing. But from a customer perspective and time to market acceleration and the ability to maintain market leadership, all of our customers are looking for utmost security and making sure that all of their secrets are maintained. So it's a really interesting time to try and maintain data in such a way that it can contribute to manufacturing acceleration or even design learnings and things like that while still maintaining all of the customer confidential information.

>> Yeah, and packaging is so important because you have a diversity of suppliers, supply chain, we talked about that, but also the partnerships are changing. So as new things come in, you want to integrate in say some sort of new silicon, I'm making this up, some sort of model, silicon, comes back. Then they said, okay, how do I test that? Okay, who's involved? Is there test cases? How do I package it in? What's the performance back to the trade-offs? These are all kind of new questions, but kind of solved in the 90s and 80s when semiconductor was booming before. Any thoughts on that kind of ecosystem partners where it's not the classic five customers or set of customers? Are there new stakeholders emerging? Anyone want to take a dealer's choice on that one?

Pam Fulton

>> I mean, I would say the memory suppliers for sure are front and center now as ... I mean they've always been part of the ecosystem, but I think that they're one of the biggest collaboration partners that all of the fabulous semiconductor companies need at this point. And same with all of us who do the manufacturing.

Nitza Basoco

>> There's other things like substrates and they're looking to embed components into the substrates. But if you do that, then how do I know that the substrate is good before I put my really expensive dye on top of it or multiple dyes on that? And so what are we doing from that perspective and innovating with the substrate vendors to ensure that they can deliver good product so that then the AAC when they get it, they can put it into the package and have faith that it's going to be great.

>> It's like onboarding a whole 'nother level of work.

Nitza Basoco

>> Yes.

>> The substrate thing is interesting. I'm glad you brought that up because I've been seeing some really conversations around substrate, the walls of a data center becoming ... Wherever there's power, give me some substrate. How much has that changed? That is a huge point.

Nitza Basoco

>> Well, it's more because power is such an issue. Can I even get power to my data center first and I only have X amount of power that I have and then, I have to distribute it? So if you come down to it, it's all about the semiconductor devices and how it's consuming it. So how can they consume less? So if you're putting the power into device to the die itself, it has to go through a packaging. So if you can minimize the power loss from the board to the die, then you're in a much better spot. This is why a lot of people are looking at substrates.

>> It's an opportunity to squeeze more and then bring an architectural element to the system.

Nitza Basoco

>> Exactly.

>> Okay. Final question. This is a little of a curveball for you guys, is that careers are changing, right? So right now you're seeing a renaissance of Semi, but it's a growth, massive growth. Careers have changed. Classic electrical engineering degree, mechanical engineering, computer science, great. Old path was kind of go in. Now you're seeing material sciences much more valuable problem solving for the folks who are learning that this is a hot area because coding is and killing the entry-level computer science jobs, I shouldn't say it that way, but changing the entry-level jobs, there's career paths into Semis. What is your guys' view on this for people watching it, looking at their careers, whether it's someone in robotics club and middle school or high school to college or even in the business, are the career paths changing? Are there some signs, signals of certain things that are in demand? Obviously, if you have this physics skills, you're going to get there, but what are your guys thoughts on this? Because I get that question a lot. I really want to jump into where the growth is, you guys are doing it. Are there areas you see that are hot and are there certain skill sets that kind of jump off the page or that it might be not well understood today that could be popular?

Pam Fulton

>> I think certainly being able to, some of us that have been around for a while maybe don't have the most experience with manipulating AI prompts and trying to understand how best to utilize the technology. So all of the people coming up now, getting the most experience they can in developing some of the AI skills and how best to leverage the technology would be great. And if I could go back and do it again, I would've paid a lot more attention to economics as well, because I think that that takes a much more powerful role than it ever has before. And I was very engineering focused through school. So if I had it to do over again, I'd pay attention to supply and demand

>> And microeconomics-

Pam Fulton

>> Yes. Yes. Yes.

>> It's math. I mean, discrete math is economics. I guess you could look at it that way. Thoughts on careers? Nitza, go ahead.

Nitza Basoco

>> Yeah, so it's so funny because I have seven godchildren and they ask me this all the time, what should I focus on when I'm in college? And it's funny because I'm an engineer and the first one is communications. Because if you have great ideas and you can't communicate them, you're not going to get anywhere. So you have to be able to communicate in a lot of different form factors, whether it'd be text, email, but in a form like this or presentation or just in meetings. But from an engineering perspective, I think electrical engineering is definitely a place to be, where a lot of people are just like, "Oh, I went computer science, computer science." But hey, you can't write your program if we don't do our jobs. And it's getting so much more complex. If you're an electrical engineer, you can go anywhere you want. EEs do a lot of a variety of different things, but also in semis, chemical engineers, material scientists, mechanical engineers, because everything is becoming multiple dimensions with these chiplets, the forces and everything there from a thermal perspective and then a physical perspectives are making things much more complex. So anything in engineering right now is a great place to be.

>> That's awesome. Okay. Just one more follow up on the market and the customers. What's the biggest change you guys will see on the market customer side from a packaging standpoint that'll be not obvious today, but maybe obvious next year?

Cherry Ooi

>> Yeah, I can add on to that. I think heterogeneous and the chip integration will be there to stay for a long time, but I think if you think of another one that will be the ... It will be still on the infancy stage, that will be the core package optic that we mentioned earlier, the CPO, judging by how fast the AI and the HPC that require the data to move so fast, CPO, you actually can make that changes. So we see that the CPO could be a game changer down the road.

Nitza Basoco

>> Great. I definitely agree with you, right? We're working with a variety of different customers to try to figure out how do we test this? And CPO is coming in and what are we doing? How do you even handle this? Do we have everything necessary across the supply chain in order to make this a reality? But it's not just co-package optics, but there's really cool new technologies like micro LED that's coming off. And from a data center's perspective, how do I change data? Can that rival co package optics? There's also RF for data transfer and what people are doing. So the startup scene is really booming right now because people are realizing, "Oh, I need hardware in order to do the things for the future and innovate."

>> I hate to say, I hate to use the Lego blocks analogy, but there's a lot ... Packaging is huge right now, right? Pam, thoughts on market customers?

Pam Fulton

>> Just, I expect everything to keep getting bigger and bigger. And I mean that in the X, Y, and Z dimension, things will get larger for the data centers in order to fit more silicon, to get more space around the outside to put all the co-packaged optics connectors and things like that. And to that end, everything needs to be known good before it gets integrated. So it's going to drive all of us to innovate in every possible dimension.

>> It's an awesome panel and it's an area that people look at and think packaging is and then, kind of make, why is that important? But I always use the example of Blackberry iPhone. It's just the packaging. I mean, it's innovative, but it's all there. How you think about it, how you test it and get it out is really awesome. Thanks for coming on. I really appreciate taking the time on the panel.

Pam Fulton

>> Thank you.

Nitza Basoco

>> Thank you.

>> I'm John Furrier here in our Palo Alto Studios for theCube and the NYSE GSA series. Thanks for watching.