Executive Vice President and Chief Technology & Products Officer, Micron Technology Inc
Scott DeBoer, Executive Vice President for Technology and Products, Micron
What role have great engineering and development played in the delivery of great products?
How does Micron plan and manage its product transitions with leadership in mind?
Let’s talk about one of those initiatives. Last year you announced that Micron would incorporate EUV into your fabs by 2024. How is that progressing?
We’ve touched on some of the DRAM nodes, but let’s dig into that roadmap a little bit. What can we expect over the next few years?
What is Micron doing to ensure technology leadership in NAND?
About Scott Deboer
In a June 2022 podcast, Scott DeBoer discussed Micron's recent progress and technology roadmap. He stated that the company's approach to product leadership involves blending a global, diverse, talented team, which he described as "the industry's best semiconductor development team." DeBoer noted that the company's performance over the prior five years exceeded his pre-COVID expectations, attributing this to a good plan, team flexibility, and remote collaboration. He highlighted a growing partnership with customers in co-inventing products, which he said helps Micron understand product needs early in the development cycle. DeBoer confirmed that Micron plans to implement extreme ultraviolet (EUV) lithography in its fabs by 2024, after a decade of work with a partner, describing the timing as critical due to added costs. He said the company will continue to drive DRAM technology leadership through planar DRAM extension and EUV integration, with a focus on products enabling Compute Express Link (CXL) and High Bandwidth Memory (HBM). Regarding NAND, DeBoer characterized the 176-layer replacement-gate technology as a "great node" that has strengthened Micron's product position, and he noted a transformation toward vertical integration across the NAND portfolio, from the transistor level through firmware and controllers. He described the current period as "a great time to be in memory" and at Micron, citing opportunities in automotive, CXL, and HBM.
Transcript (24 segments)
P
Pat Morehead0:10
Micron EVP of Technology and Products, Scott, how are you?
S
Scott Deboer0:15
Great, great to be here today, Pat. Really appreciate the opportunity.
P
Pat Morehead0:18
Absolutely. You know, you really had a rocking investor day. It had been a few years since you had one. I think everybody was really happy that you came and gave a really high degree of detail. But you know what, we're going to do a little bit more of a double click here on what you brought out. Where I'd love to start off is essentially, what is your approach for product leadership? It's clear to me as an industry analyst, and I've been in this industry over 30 years...
S
Scott Deboer1:11
Yeah, thanks. You know, I've been in the industry a long time too, and we've certainly had different approaches through time and different degrees of success. I think one of the great things about this industry event, analyst event really, was we had the opportunity to tell the story of how much progress we've had over the past five years. That really is a result of our approach of blending a global, diverse, talented team. I mentioned then, and I truly do believe this, this is the industry's best semiconductor development team. I love it. It's come together across multiple parts around the world.
P
Pat Morehead2:12
10 years evaluating semiconductors and I guess 11 years as an analyst, so I can't even imagine putting together and engineering and then fabbing a design during COVID. But it blows my mind. I mean, was that a challenge for you all?
S
Scott Deboer2:33
Well, it's a great question, Pat. I think I will agree with you. Going into COVID, I would have thought that many of the things that we did and were able to do likely were impossible. We certainly had a good plan and a great team, and I think it reflects on the flexibility of that team and the partnerships... which really was just starting at Micron at the beginning of COVID. It's amazing what was able to get done, but it did require the team to focus on new ways of doing things through this period of time.
P
Pat Morehead3:28
You know, one way, it's funny sometimes people take this for granted, but I don't because I worked at a leading microprocessor company for 11 years. The magic when you get bring up with a new memory technology, and I think a lot of that has to do with what I've seen from you: heavy-duty partnerships and, dare I say, co-invention with some of your customers. Am I looking too much into this, or do you really... helping us understand from the early parts in our development cycle what those products need?
S
Scott Deboer4:12
And that is a unique thing. I think it's partially driven by the recognition across the globe, but our customer base for sure, of the importance of memory and the growing importance of memory, and our ability to work with them. Of course, once you execute and you deliver some things for customers that provide them unique value, then it just kind of grows and becomes a way of doing business. I think that's what's happening now.
P
Pat Morehead4:48
What amazes me is when a new memory technology and a new processor host show at the same time... limitations for not having that planning done. I remember the DDR, DDR2 days and crazy stuff. So, one of the things that I've learned in my travels is that the winners and losers sometimes come with product transitions. I'm curious, this is another approach question for you: how do you look at product transition and keep the leadership going? I mean, in one way I could say you're a leader in some of the most recent product transitions.
S
Scott Deboer5:51
Yeah, it's another really good one. The time over the last... some transitions that are more operational, which was related to our improvement in cadence in terms of how fast we roll out technology nodes over a period of time to get more competitive. We've really looked at those as the opportunities where we want to win and make the right decisions. At times, those are not obvious to others in the industry or analysts why we're doing specific things, but I think we're very proud of how a number of these have played out for us, whether it's our transition from floating gate to replacement gate NAND, which was frankly a pretty big bet when we did it.
P
Pat Morehead7:10
Some questions I was taking at that time: 'Okay, well, are you really going to change memory technology and come out in any place other than lagging the industry in technology?' And I think we're very proud of how we came out of that, not just with strong products but really a technology lead of a year or more and some really strong end products based on that technology. So that's a good one, and there's others that we're proud of as well.
S
Scott Deboer7:40
Yeah, so does this only apply to the here and now, or does this also apply to the future? I mean, there are a few product transitions and technology transitions coming up in the future here. Hopefully, you're going to run that same play there.
P
Pat Morehead8:11
Confidence and a demonstrated ability to prepare for these big ones coming at us in the future. Certainly, there are product transitions in terms of architecture: AI driving different kinds of memory applications, architecture changes coming up at some point on the DRAM side, and next steps of UE. So there are some big ones coming up, and we certainly intend to win as we go through those transitions too.
S
Scott Deboer8:43
Yeah, two that I'm really excited about: next generation HBM. I mean, we're all realizing that monolithic designs are... they're re-architecting their hyperscaler data centers to get ready for that next wave. But also from a technology perspective, EUV is a biggie. You know, it's funny, I saw companies maybe six or seven years ago saying, 'We don't know if the value is in EUV, we're going to wait,' and they got hammered. But last year you announced that you would incorporate EUV into your fabs by 2024. How's that working out?
Sure. That is one of the transitions that I feel quite pleased with how we've managed. We stayed very... late, and I think 2024 is the time when we're going to be able to implement an infrastructure around EUV that is very capable for high volume manufacturing. Based on our leadership on our 1-alpha and 1-beta DRAM, I do believe that's partially because of us making the right call on how long to implement multiple patterning technology and when to bring EUV in.
P
Pat Morehead10:40
Yes, this is not something you want to take lightly, especially when you look at the added cost that it brings to the table. So it has to be the right time for you, and it does help remove some of the future barriers to DRAM over the next few years.
S
Scott Deboer11:12
Sure. I think we'll continue to drive our DRAM technology to ensure that we have industry leadership in terms of cost structure and performance. Certainly, that will be based on a planar DRAM extension and driving EUV into the process flow, as you brought up rightly a minute ago, and driving multiple more nodes on the planar DRAM side. More and more focus on DRAM that enables things like CXL and HBM that you mentioned. Those have significant growing importance, and we intend to be in a strong position with those advanced products and have the DRAM products to support it.
P
Pat Morehead12:10
Some of the critical ones will be EUV, and the most critical ones will be EUV, and then we'll still blend in a mixture of multiple patterning. So across the whole process flow, we'll have a mix of optimal times when we use EUV and when we don't. Do you still feel, put you on the spot here a little bit, do you still think you'll have a competitive advantage here?
S
Scott Deboer12:36
Yeah, we're very comfortable in driving our DRAM roadmap. I mentioned our 1-beta node coming, we'll have production on before the end of the calendar year, and we believe that will be a leadership position again. So that'll be for the next couple years. Time will tell on those.
P
Pat Morehead13:10
Yeah, what a market. I mean, I do a lot of market scoping, and the increase in bits in markets like automotive and compute is absolutely mind-blowing. I think if we could have gotten enough graphics chips to match with memory, we had the memory, we didn't have enough GPUs. I think we would have seen a lot more memory put on those GPUs. In fact, they actually had to pull back on some of the memory per graphics card to be able to ship lower priced configurations because the prices of those graphics cards just got crazy. The good news... had some people who weren't shipping with radios because they couldn't get the analog bits they needed. But one thing I know for sure is that in automotive, as we increase electrification, we have standard L2 ADAS and we move into autonomous driving, you're going to see a lot more memory show up in these cars. Because quite frankly, when you look at even ADAS, I'm not going to say L2 requires a supercomputer on wheels, but for autonomous driving you do need gigaflop capabilities with a lot of memory for real-time reaction. So hey, we've given a lot of love to memory here, but let's go into storage.
S
Scott Deboer15:10
Replacement gate technology that we were discussing earlier, that is a great node for us. The volume of our bits... it is a solid platform that has really led to our ability to focus on much better end system products. Over the past five years, another big transformation for us has been in really having vertical integration across our NAND product portfolio in terms of optimizing our technology from the transistor level through the end system, the firmware, the IP that we put into the ASIC in our controllers, and really optimizing those. We're starting to see a much stronger product position. Really proud of how our team has put together a NAND product portfolio now that is the equal of the NAND technology that we've been able to do, and combine those two.
P
Pat Morehead16:28
Yeah, it's like firing on all cylinders here. No listen, I've been amazed particularly with QLC, what some of your customers have been able to do with that in terms of reliability. It's amazing. It's so funny, I think I've heard for the last 10 years that spinning disks are going to go away, and QLC is the next iteration of that. I think the great thing is that the entire industry keeps raising the bar in the entire storage tier. I call it the Arctic freeze. Systems are just creating so much data on this planet. You and I have both seen all the slides and we don't want to bore people with that, but it is amazing that we keep increasing by order of magnitude the amount of data that needs to be stored. Just like we saw in DRAM, it's data center, it's PC, it's automotive. I mean, we have terabyte on smartphones, which might not be sexy to somebody who's in a big desktop, but it's absolutely incredible. You can't... need some longer term, but low latency storage, you're going to put it on an SSD. So I don't see any increase in demand for NAND in the foreseeable future; I just literally see order of magnitude. So it's got to be an exciting place to be here at Micron. And maybe, Scott, we'll kind of end on you. Is there anything that I didn't ask that you wanted me to ask, or any final words?
S
Scott Deboer18:46
Well, I think you rightly stated it's a great time to be in memory, great time to be at Micron. The opportunities ahead are really phenomenal. You touched on... exciting piece for us. Certainly you touched on CXL and HBM, and we are very focused and excited to be creating products there. It really is a place where that co-invention with customers is kind of highlighted, because those are the leading edge of how next-generation computers are really going to work, and memory is a very key part of it. So we're really excited about the future, and I'm definitely happy to have the opportunity today to chat with you.
P
Pat Morehead19:41
Yeah, I appreciate that, Scott. And listen, everybody likes a success story. I think to take five seconds and pat yourself and the team on the back. I've been monitoring it, I wasn't quite sure how you were... gotta make sure this market keeps going here. But thanks, Scott, so much for taking the time and adding the value to the audience for the More Insights and Strategy podcasts. Thank you. And with that, this is Pat Morehead signing off. If you like what you heard, hit that subscribe button. Have any comments for myself, you know where to find me on Twitter. Complaints and accolades are welcome. So with that...