ADVANCED MICRO DEVICES INC — Call Transcript 2026

Back to this BofA Global Technology Conference. I'm Vivek Arya from BofA's Semiconductor, Semi Cap Equipment research team. I'm really delighted to have the team from Advanced Micro Devices join us this morning, Jean Hu, Chief Financial Officer, and Matt Ramsay, Head of Investor Relations. I'll go through my questions, but please feel free to raise your hand if you would like me to bring up something. Really warm welcome to you, Jean and Matt. Really appreciate you joining us - Yeah. Thank you. - at the conference. I guess exciting times in AI and the semiconductor industry. Maybe Jean, as a start, just walk us through your state of the union, right? The growth has been very strong, right, since the start of the year. What has surprised you? Which are the areas that you think you are seeing the best growth right now, and which are the areas which have been pushed back because of this? Yeah. First, thank you for having us. Thank you all for joining us. I think the biggest change during the last few months is really the rise and inflection of agentic AI. Right. You can see continued momentum from training to inference, from AI adoption experimentation to more adoption at scale. When you look at that, agentic AI, it's not about answering question anymore. It's about orchestration, it's about database access, and a lot of tool execution. All of those require significant CPU performance. What we're seeing is very significant and the incremental demand for our CPU platforms. That has been really exciting. At the same time, we also are seeing economics of AI keep changing, right? With the token generation going up so quickly, all the customers are really focusing on performance, TCO, and they are really trying to figure out how you use different compute to address different application and the workload. We see demand for GPU going up, CPU, and also a lot of other ASIC LPUs. I think from AMD's perspective, you know our Q1 performance. Right. We had a record CPU performance. CPU business grow more than 50%. We guided the Q2 CPU performance is going to go up year-over-year 70%. That has been really exciting. All those things really benefit us because we have been investing in GPU, CPU, adaptive compute to also address literally end-to-end applications from data center to client to gaming. That feels like right now all the engines are really driving the business revenue growth, so very significantly. Got it. I think the growth in CPU has been kind of the biggest positive drivers this year. We are kind of tracking this 2030 or late end of the decade type of addressable opportunity, which I think you were the first to point out would be $60 billion. Arm said, "That doesn't sound that high enough." They took it up to $100, then you took it up to $120, and Jensen said, "Well, could be even more than that," right, up to $200. What do you think is the difference between all these forecasts? This is one of the most frequently asked questions from the investment community. What is the right number? How does one get to it? Is there like a simple unit, and are we just all using different ASPs to get to a different number, and everyone has the same unit growth? You're asking a lot of questions here. I think maybe let's take a step back. You are absolutely right. AMD was the first one. Literally last year, we start to talk about how AI drive the demand of CPUs. Last November when we had our financial analyst day, we actually outlined how we think about this opportunity under the TAM. We basically said, "Okay, we are thinking three segments." One is the traditional general compute, which we are all very familiar with, has been driving the traditional enterprise application across the board. The second category is the head node, which is really the communication with the GPUs. You want speed to make sure you have the communication. We also outlined agentic AI last November, which is still very early, but our team already see the early signs for agentic AI. That has been the fastest-growing market in our view. That's when we said, "Okay, we believe the TAM is going to grow 18% CAGR to $60 billion in 2030." Remember, the general purpose CPU used to be just a single-digit growth. Right. That market expansion, we saw it earlier, but we never estimate it's going to be so fast. The pace and the speed of agentic AI adoption has been tremendous. When you look at the January, February, March, all the enterprise adoption started. You see Anthropic, OpenAI, their revenue has been going up very significantly. The demand for CPU continued to go up across all our customers. That's when we had our earnings call. We actually updated our market forecast to more than $120 billion. It's still very early when you think about agentic AI adoption. What we are seeing is with agentic AI, it's quite complex when you have millions of operations you have to execute. When you have a lot of agent concurrently working on different tasks, you actually need a very high core count CPUs. We don't know what other companies, how they think about the TAM and the SAM. We go through the bottom-up and the top-down approach. We know at this particular point, we feel pretty good about more than $120 billion, but it's evolving so fast, and the more complex you should expect for the agentic AI portion of the market, because you have so many diverse workloads, so complex, the ASP will continue to increase because increasingly you need a very high core count CPUs, high performance CPUs. Matt, I don't know if you have other you can add here. Sure. First off, thank you, Vivek, for hosting us and for you guys all coming. I think there's a couple things, Vivek, that I would maybe add to Jean's comments. One is, I think for this whole audience and us included, we were trying to figure out when was the primary driver of AI CapEx going to go from almost predominantly large model training and start the shift towards inference being a primary driver of CapEx. We're seeing that happen in real time. I think the more powerful thing for our business and maybe the more transformative thing architecturally is during that transition, you're seeing chatbot inference become agentic inference, and what agentic inference does as a workload. You daisy chain or concatenate all these automated agentic flows together, between each inference task, there's a lot of CPU diverse work, so post-processing of data, figuring out what to tell the AI to do next, often based on the result of the prior inference. Where do I get the data from for the next step? Is it in a cloud? Is it in an ERP system, a payment system, a CRM system, wherever? Get all of that data, come back, post-process the data and from that step, and feed it into the next agentic step for the AI. That workload is quite diverse. That's what we're seeing driving the demand for Turin, now in real time. As the order book fills in for the 256 core 2 nm Venice parts, that are going to launch in a couple of months and be the primary workhorse for next year. That's why we're seeing the order book really expand. Got it. Is there a way to segment that market, whatever it happens to be? $120 billion+, in those three segments, Jean, that you described, the kind of the traditional applications, let's call it part of the AI cluster and then agentic kind of standalone CPUs? I think a traditional segment, it's very clear. I think 2025, a different third party would say that's a $25 billion-$30 billion market opportunity. It will continue to grow. As Matt has mentioned, is you actually have an agent doing more work on the traditional database. Right. That traditional general purpose CPU, it's probably steady growth, but not as high. The head node, which is working with the GPUs, you can see that's when people talk about the ratios, is traditionally it's one head node CPU manager, eight GPUs - Nice. - two to, yeah. It's getting more and more. Over time, that ratio is changing. We do think the head node segment CPU will also grow very fast, much faster than the traditional general CPU. The most exciting portion of the market is actually agentic AI. Agentic AI, you actually are seeing agentic AI server rack sit in between the traditional servers and the GPUs. Those racks are handling all those different workload to really make sure all the agentic agent works. That market, we think by the whatever, the $120 billion or $200 billion market opportunity is majority of that large market. Over 50% of - Over 50%. - that market you think is that standalone? Agentic. Yes. Yeah. It's still very early right now, but if you just think about what Matt said, how complex, how dynamic those workload will be, and we do see significant productivity improvement, especially software engineering side. Nice. Those are very complex workload. You really need different core counts. You need really high performance, high core counts. You also need simultaneous multi-thread, all those high performance, you can manage millions of agent potentially. Got it. The final question there is actually two last questions. One is x86 versus Arm. If let's say that agentic opportunity does capture the dominant share of whatever that TAM number happens to be, what advantage, and if I dare ask, what disadvantage does x86 have versus Arm in capturing that? I'll start. Matt can talk about the more technical side is the way AMD has been thinking about it has never been, oh, x86 architecture or Arm architecture. It's more about how we provide the best TCO for customers. For us, the performance is number one. Secondly, what we have been trying to build is the breadth and the depth. For platform portfolio. If you look at our CPU platform, we have gone through five generation at the Turin and the Venice is next generation. We have the breadth from the core count to 8 to 16 core count addressing enterprise applications to the Venice will have 256 core count. For us, we have all different configuration design points to meet all the different compute to give them best TCOs. I don't know technically you can provide them. Yeah. Vivek, I think as Jean talked about, the three different buckets, right? The traditional server workloads for enterprise, I think there is an affinity towards x86 because of the code base. Many of those servers are deployed on-prem, some of them are deployed in third-party cloud. It's basically the same workload depending on where you want to deploy. I think as these agents generate more traffic to all those traditional systems, the x86 ecosystem is well positioned there. You have the head nodes. I think that what you really need in head nodes is really high single-thread performance, really high bandwidth to, I don't know, it's not a technical term, but feed the beast of these really expensive accelerators that are in the box, right? We will do our own head nodes for Helios. I think our large competitor in their NVL racks will basically do their own head nodes. There's a bunch of different XPU programs that are going to be launched by the industry. I think when we have those conversations with customers, it's much more about what is the performance of your CPU such that I get the best utilization out of this really expensive XPU that I'm buying. It's not an Intel or AMD or x86 versus Arm conversation. It's a what's the best server part. When you talk about these agentic racks, it's really about how many concurrent agents can you run per rack or per megawatt of compute. As Jean mentioned, really high thread and core count products. I think our roadmap is differentiated there. Another thing that doesn't get mentioned, but I think is important, is if you think about the dollars of CapEx that are going to come into the server market, the mission-critical nature of these systems, both head node and agentic racks. Some of the RAS features and the security features of the x86 ecosystem, it's not an x86 architecture thing, it's just we and our x86 competitor have been put through the paces by every hyperscaler and most enterprises on all of these security features for a decade. You might imagine, you get into these automated flows where those agents have access to mission-critical data, and you want to really have servers that have robust security features there. We feel we get brought into essentially every RFQ that's in the industry for these servers, and I think we're positioned to win a very high percentage of them. Got it. Now, your x86 competitor has their own fab. They are planning to add a lot more capacity on the server CPU side. I guess this year every CPU is being sold, right? Do you think that there is a scope for share shifts given that your competitor has incremental capacity to devote to this, or you think you will be adequately served by your foundry ecosystem? Yeah. We have been very pleased with our share gain trajectory. I think the last Q1, which is announced, we actually from a value share perspective, we got to like a 46%. Nice. When you think about it, our success is really tied together with TSMC. From day one, we have been using TSMC. We work together with our chiplet design, with the packaging technology to come up with the best performance server platforms. That continue to benefit us going forward. We have been planning for this ramp since last year. If you look at the Q1 performance and the Q2 guidance, when we guide year-over-year 70% increase. A lot of you actually know is right, the way we first started six or nine months ago. It is the early planning, how we work with the TSMC to make sure we get support to continue to drive the ramp of the supply. The demand is tremendous. Supply is still tight. I think for us, for this year, next year, every supply we can get, we can provide to our customers. We're going to continue to do that. It's the planning beyond the 2027 right now. For the longer term, I think we'll continue to work with TSMC and prioritize data center as how we drive the business growth going forward. Got it. Vivek, one little thing that I would add to what Jean mentioned there is when we have conversations with the investor community on this particular topic, I think one thing that I've noticed is supply is tight, right? 3 nm tight. There's a lot of other areas that are tight, and I think we're very well-positioned with the relationships Lisa has personally with the executives in that space to get maybe more than our fair share of the incremental. It is tight. I think what I've noticed is maybe the investor scoping of what Lisa and the team initially asked for in 2026 and 2027 maybe didn't imagine how much growth we had already planned for in supply. Things are tight, but we've already asked for and already been sort of, quote-unquote, "granted" from TSMC, I think we're having the right conversation there, but maybe not at the right starting point from what we were sort of allocated initially. Got it. Final one on the CPU side on price. You mentioned 50%+ growth in Q1 and then 70%+. I think there was a sense that one could sustain this - Yeah. - for the rest of the year. How much of this, Jean, is pricing versus unit growth? Let me just add quickly part B to that. As that agentic, that third segment, becomes a bigger portion, do you think that can command much higher ASPs? When I talk with the Arm camp, they're talking about ASPs in the $3,000, $4,000, $5,000 range, which is well above the ASPs that AMD is able to get. Just unit versus ASP and just the trend of ASP. Yeah. I think as you mentioned, both the more than 50% growth in Q1 and that we guide the Q2 70% year-over-year, more than 70% year-over-year increase, two-third of that actually are unit growth. Majority of our revenue growth actually come from unit expansion versus ASP increase. Your point on the ASP going forward, especially with the agentic AI, is a great point. What we're seeing is we're at the early inning of agentic AI adoption. You can already see with the software engineering part of agentic work, autonomous agentic workflow, you really need a very high performance, high core counts, CPUs. Right. That tend to have a higher ASP. I think a generation over generation, when we go to 2 nm with Venice and the future generation to handle complex workload to provide the performance, you tend to see the CPU price continue to increase. The other thing I would just say is because we have a really large platform, the CPU ASP is really determined by the mix - Right. - by the different configurations for different workload. For enterprise, the per core ASP is very high, but because the core counts are relatively low, so ASP is relatively low, but the gross margin is great. It's very nuanced from our business because how big our platform is and how complex all different design point we're helping our customers. Got it. You spent two thirds of the time on CPU, let's talk about GPU. It's great. GPU now. You are at the start of launch of a major rack-scale system to your portfolio. Maybe give us a status update of how you're feeling. When your competitor launched their first large scale up product, it took a while for the industry to get ready, liquid cooling and all those things to get ready. How do you think about your progress so far? How excited are you about the second half? Yeah, we are on track. We sampled MI450. We expect to be launching in second half or Q3 starting point and the Q4. As you know, we have been preparing for this launch for a long time. We acquired ZT Systems in 2024. The whole reason we're acquiring ZT Systems is to help us to add capabilities in system-level design. We also have worked on the networking side. The whole team has been working to prepare this launch. We know it's very hard. It's very complex, very hard, but from execution perspective, our team has been on track, meticulous prepare for all different aspect launch. Not only technology software side, but more importantly, the supply chain side, how you make sure the design redundancy with all the components, but also supplies even very small components, you have to make sure you have supplies. That has been the whole team's execution track record. We feel pretty good right now. Vivek, I think we're anticipating, as Jean said, we already have sampled and we have a number of customers that have full Helios racks in their own data center running their full production workload now and testing and doing all those things that you would expect us to be doing at this point in time. The customers learned a lot, we learned a lot from some folks that blazed the trail, as you mentioned. We're going to really focus on a concentrated set of ODMs as we launch for the first number of months until we get up to scale, not like a few systems, but billions of dollars of scale. We'll be expanding the ODM ecosystem pretty significantly as we move through next year and get towards the MI500 series and whatnot. We feel good about where we are. The initial work that the customer base is doing in their own labs with sampled systems has gone very well. Now it's about We'll see a pretty significant jump in the fourth quarter of revenue, another fairly significant jump in Q1, and then we'll go from there. It is wood, so touch wood. Everything's going exactly as we'd hoped it goes. When you're doing a thing like this, day to day, there's always something. Our job's to keep the duck calm on the top as we kick on the bottom and get ready to ramp this thing starting in Q3. Got it. When you've been asked, Jean, or Lisa has been asked, you have always mentioned the opportunity for multi-gigawatt scale deployments. I think we have heard about two of them, OpenAI and Meta. You think there is still the scope for us to be pleasantly surprised with more such announcements even for 2027? You think that just given how tight the industry is, that more incremental announcements could be further spaced out than that? Yeah. We talked about we're really pleased with our partnership, very long-term partnership with both OpenAI and Meta. Last year, we established those relationship, and we actually see the forecast from those customers are actually above our original plan for 2027. I think one of the key milestone we're really focusing on this year is to continue to expand and deepening the customer adoption of MI450. We definitely expect there will be other multi-gigawatt customers. We have both, across the board, hyperscale customers, the model companies, and even AI native companies who we have been engaging, working with. I think it's important for us to continue to expand this customer base and with the largest scale deployment. Got it. Makes sense. How do you expect the two customers you have announced to allocate market share? Because they have a number of supplier options. I know you have a somewhat unique engagement with them through the granting of warrants and whatnot, but just if we set that aside, how is your visibility into getting, I think the two gigawatt each from those customers for next year? Vivek, as you know, the lead time for MI450 is quite long. We really have to start to plan and have the orders from customers for next year for the second half launch and the next year. I think the supply chain overall is tight next year. Since we have been preparing to support our customers, we do want to ensure we get the supplies for what they need. Those are two our most important anchor customers. Of course, we also have a warrant, which we're going to be able to share the upside if we get more revenue. Our stock price is higher, they will share upside too. That is absolutely the common interest for both companies who want to drive the adoption of MI450. We feel pretty good with both of them. We think we can get the right supplies to support them. Got it. Do you think having someone like TSMC is almost adding a level of discipline in the ecosystem? It's just harder to double count if everyone has to go to the same source. I know it sounds obvious, but do you see that also? They're quite disciplined. I think they'll continue to be. These are sensitive topics for us to talk about in these forums, but we really, really, really appreciate the support that TSMC giving us as a partner, and the way that they're working with us to bring more capacity online to support not just the Helios and 450 growth, but the significant server TAM expansion that we're seeing and planning for not just for this year, but for all of next year and conversations going into 2028 on wafer capacity to support demand. We really appreciate the work that they're doing on our behalf, and I think they're being, as they've always been, disciplined in the way that they bring capacity online and do their diligence two or three levels deep in the customer base to understand where the organic demand is coming from to support the capacity that they're bringing on. Makes sense. The final one, Jean, is on just memory cost inflation. I guess, when you wake up and see DRAM, I am sure DRAM price is not what you see when you wake up, but when you see that cost inflation every day, is that a good thing because it gives you more pricing power, or is that a bad thing because now you have to absorb that cost and hunt for more supply? What are kind of the pros and cons of that memory cost inflation for you? It's really a great question. First for us, we really want to make sure if there's a longer-term agreement with our suppliers, we can secure the capacity. That is number one thing when you have a larger business at our scale, you have to plan ahead. Secondly, we think more strategically, if there's memory cost increase, we do need to figure out a way to share with our customers to absorb that increase together. I think the most important thing for us is we actually don't want to use the shortage of controlled memory to just increase the price. We want to make sure customer help us to absorb the cost increase, we are more strategic with the long-term relationship. I think from industry perspective, the memory cost increase at this kind of level, we have never seen. The price is really high. I think it does increase the CapEx spending for end customers where they spend the CapEx. I think in the longer term, just like all the industry, you always have a new capacity add up. You also have a new innovation on how you utilize memory more efficiently. It will figure out itself in the longer term. That's what we believe. Got it. Do you have long-term agreements in place for the majority of your memory needs, or do you think that this is a cyclical high, and maybe there will be more normalization of that input cost for you? From planning perspective, because our manufacturing cycle is so long, if we plan for the MI450 ramp this year, we had to start to talk to our suppliers last year or year before. Similar situation is right now, not only with the very tight supply environment, not only we're planning 2027, we're planning 2028. For us to continue to make sure we can support our ambitious plan to grow the company, we have to plan ahead. Yeah. Vivek, in the server market and in the PC market, the vast majority of the DRAM is actually acquired by our OEM and ODM partners and the hyperscalers themselves. That DRAM doesn't run through our P&L, and we work really, really hard with all of the customer base to try to make sure that we're matching the memory supply that they're able to get and the CPU supply or the GPU supply that we're able to provide together to where there's not mismatching of supply in the industry or hoarding of anything. We've been working really hard on that. There's certainly increased input costs to end devices in certain parts of the consumer and gaming market where there have been some anticipated market impacts, and I think we've given you guys our current view of that and the guidance that we've talked about. We're working really hard on behalf of the whole customer base to try to make sure that the memory pricing and the matching of components is relatively seamless in a time like this. Got you. Just final question before we close. When you look across your customer base, Jean, how far is the visibility that they're willing to provide to you now, and when you get that, what are you doing differently in your business operationally to kind of deliver to those? Yeah. I think because the supply chain environment is really tight, customer really plan ahead. We have a very good visibility into 2027 and beyond. I think our customers and ourselves, especially when you think about the data center investment, very large CapEx planning, very long planning cycle, they actually plan way ahead. 2027, we feel really good about both the visibility of demand side and the visibility of supply side. Of course, 2028 and beyond, that's everybody's working on. Okay. On that positive note, thank you so much, Jean. Thank you. I appreciate your time.

Speaker 3: Back to this BofA Global Technology Conference. I'm Vivek Arya from BofA's Semiconductor, Semi Cap Equipment research team. I'm really delighted to have the team from Advanced Micro Devices join us this morning, Jean Hu, Chief Financial Officer, and Matt Ramsay, Head of Investor Relations. I'll go through my questions, but please feel free to raise your hand if you would like me to bring up something. Really warm welcome to you, Jean and Matt. Really appreciate you joining us - Back to this BofA Global Technology Conference. back to this bofa global technology conference I'm Vivek Arya from BofA's Semiconductor, Semi Cap Equipment research team. i'm vivek arya from bofa's semiconductor semi cap equipment research team I'm really delighted to have the team from Advanced Micro Devices join us this morning, Jean Hu, Chief Financial Officer, and Matt Ramsay, Head of Investor Relations. i'm really delighted to have the team from advanced micro devices join us this morning jean hu chief financial officer and matt ramsay head of investor relations I'll go through my questions, but please feel free to raise your hand if you would like me to bring up something. i'll go through my questions but please feel free to raise your hand if you would like me to bring up something Really warm welcome to you, Jean and Matt. really warm welcome to you jean and matt Really appreciate you joining us - really appreciate you joining us -

Speaker 1: Yeah. Thank you. Yeah. yeah Thank you. thank you

Speaker 3: - at the conference. I guess exciting times in AI and the semiconductor industry. Maybe Jean, as a start, just walk us through your state of the union, right? The growth has been very strong, right, since the start of the year. What has surprised you? Which are the areas that you think you are seeing the best growth right now, and which are the areas which have been pushed back because of this? - at the conference. - at the conference I guess exciting times in AI and the semiconductor industry. i guess exciting times in ai and the semiconductor industry Maybe Jean, as a start, just walk us through your state of the union, right? maybe jean as a start just walk us through your state of the union right The growth has been very strong, right, since the start of the year. the growth has been very strong right since the start of the year What has surprised you? what has surprised you Which are the areas that you think you are seeing the best growth right now, and which are the areas which have been pushed back because of this? which are the areas that you think you are seeing the best growth right now and which are the areas which have been pushed back because of this

Speaker 1: Yeah. First, thank you for having us. Thank you all for joining us. I think the biggest change during the last few months is really the rise and inflection of agentic AI. Yeah. yeah First, thank you for having us. first thank you for having us Thank you all for joining us. thank you all for joining us I think the biggest change during the last few months is really the rise and inflection of agentic AI. i think the biggest change during the last few months is really the rise and inflection of agentic ai

Speaker 3: Right. Right. right

Speaker 1: You can see continued momentum from training to inference, from AI adoption experimentation to more adoption at scale. When you look at that, agentic AI, it's not about answering question anymore. It's about orchestration, it's about database access, and a lot of tool execution. All of those require significant CPU performance. What we're seeing is very significant and the incremental demand for our CPU platforms. That has been really exciting. At the same time, we also are seeing economics of AI keep changing, right? With the token generation going up so quickly, all the customers are really focusing on performance, TCO, and they are really trying to figure out how you use different compute to address different application and the workload. We see demand for GPU going up, CPU, and also a lot of other ASIC LPUs. I think from AMD's perspective, you know our Q1 performance. You can see continued momentum from training to inference, from AI adoption experimentation to more adoption at scale. you can see continued momentum from training to inference from ai adoption experimentation to more adoption at scale When you look at that, agentic AI, it's not about answering question anymore. when you look at that agentic ai it's not about answering question anymore It's about orchestration, it's about database access, and a lot of tool execution. it's about orchestration it's about database access and a lot of tool execution All of those require significant CPU performance. all of those require significant cpu performance What we're seeing is very significant and the incremental demand for our CPU platforms. what we're seeing is very significant and the incremental demand for our cpu platforms That has been really exciting. that has been really exciting At the same time, we also are seeing economics of AI keep changing, right? at the same time we also are seeing economics of ai keep changing right With the token generation going up so quickly, all the customers are really focusing on performance, TCO, and they are really trying to figure out how you use different compute to address different application and the workload. with the token generation going up so quickly all the customers are really focusing on performance tco and they are really trying to figure out how you use different compute to address different application and the workload We see demand for GPU going up, CPU, and also a lot of other ASIC LPUs. we see demand for gpu going up cpu and also a lot of other asic lpus I think from AMD's perspective, you know our Q1 performance. i think from amd's perspective you know our q1 performance

Speaker 3: Right. Right. right

Speaker 1: We had a record CPU performance. CPU business grow more than 50%. We guided the Q2 CPU performance is going to go up year-over-year 70%. That has been really exciting. All those things really benefit us because we have been investing in GPU, CPU, adaptive compute to also address literally end-to-end applications from data center to client to gaming. That feels like right now all the engines are really driving the business revenue growth, so very significantly. We had a record CPU performance. we had a record cpu performance CPU business grow more than 50%. cpu business grow more than 50% We guided the Q2 CPU performance is going to go up year-over-year 70%. we guided the q2 cpu performance is going to go up year-over-year 70% That has been really exciting. that has been really exciting All those things really benefit us because we have been investing in GPU, CPU, adaptive compute to also address literally end-to-end applications from data center to client to gaming. all those things really benefit us because we have been investing in gpu cpu adaptive compute to also address literally end-to-end applications from data center to client to gaming That feels like right now all the engines are really driving the business revenue growth, so very significantly. that feels like right now all the engines are really driving the business revenue growth so very significantly

Speaker 3: Got it. I think the growth in CPU has been kind of the biggest positive drivers this year. We are kind of tracking this 2030 or late end of the decade type of addressable opportunity, which I think you were the first to point out would be $60 billion. Arm said, "That doesn't sound that high enough." They took it up to $100, then you took it up to $120, and Jensen said, "Well, could be even more than that," right, up to $200. What do you think is the difference between all these forecasts? This is one of the most frequently asked questions from the investment community. What is the right number? How does one get to it? Is there like a simple unit, and are we just all using different ASPs to get to a different number, and everyone has the same unit growth? Got it. got it I think the growth in CPU has been kind of the biggest positive drivers this year. i think the growth in cpu has been kind of the biggest positive drivers this year We are kind of tracking this 2030 or late end of the decade type of addressable opportunity, which I think you were the first to point out would be $60 billion. we are kind of tracking this 2030 or late end of the decade type of addressable opportunity which i think you were the first to point out would be $60 billion Arm said, "That doesn't sound that high enough." They took it up to $100, then you took it up to $120, and Jensen said, "Well, could be even more than that," right, up to $200. arm said "that doesn't sound that high enough." they took it up to $100 then you took it up to $120 and jensen said "well could be even more than that," right up to $200 What do you think is the difference between all these forecasts? what do you think is the difference between all these forecasts This is one of the most frequently asked questions from the investment community. this is one of the most frequently asked questions from the investment community What is the right number? what is the right number How does one get to it? how does one get to it Is there like a simple unit, and are we just all using different ASPs to get to a different number, and everyone has the same unit growth? is there like a simple unit and are we just all using different asps to get to a different number and everyone has the same unit growth

Speaker 1: You're asking a lot of questions here. I think maybe let's take a step back. You are absolutely right. AMD was the first one. Literally last year, we start to talk about how AI drive the demand of CPUs. Last November when we had our financial analyst day, we actually outlined how we think about this opportunity under the TAM. We basically said, "Okay, we are thinking three segments." One is the traditional general compute, which we are all very familiar with, has been driving the traditional enterprise application across the board. The second category is the head node, which is really the communication with the GPUs. You want speed to make sure you have the communication. We also outlined agentic AI last November, which is still very early, but our team already see the early signs for agentic AI. You're asking a lot of questions here. you're asking a lot of questions here I think maybe let's take a step back. i think maybe let's take a step back You are absolutely right. you are absolutely right AMD was the first one. amd was the first one Literally last year, we start to talk about how AI drive the demand of CPUs. literally last year we start to talk about how ai drive the demand of cpus Last November when we had our financial analyst day, we actually outlined how we think about this opportunity under the TAM. last november when we had our financial analyst day we actually outlined how we think about this opportunity under the tam We basically said, "Okay, we are thinking three segments." One is the traditional general compute, which we are all very familiar with, has been driving the traditional enterprise application across the board. we basically said "okay we are thinking three segments." one is the traditional general compute which we are all very familiar with has been driving the traditional enterprise application across the board The second category is the head node, which is really the communication with the GPUs. the second category is the head node which is really the communication with the gpus You want speed to make sure you have the communication. you want speed to make sure you have the communication We also outlined agentic AI last November, which is still very early, but our team already see the early signs for agentic AI. we also outlined agentic ai last november which is still very early but our team already see the early signs for agentic ai That has been the fastest-growing market in our view. That's when we said, "Okay, we believe the TAM is going to grow 18% CAGR to $60 billion in 2030." Remember, the general purpose CPU used to be just a single-digit growth. That has been the fastest-growing market in our view. that has been the fastest-growing market in our view That's when we said, "Okay, we believe the TAM is going to grow 18% CAGR to $60 billion in 2030." Remember, the general purpose CPU used to be just a single-digit growth. that's when we said "okay we believe the tam is going to grow 18% cagr to $60 billion in 2030." remember the general purpose cpu used to be just a single-digit growth

Speaker 3: Right. Right. right

Speaker 1: That market expansion, we saw it earlier, but we never estimate it's going to be so fast. The pace and the speed of agentic AI adoption has been tremendous. When you look at the January, February, March, all the enterprise adoption started. You see Anthropic, OpenAI, their revenue has been going up very significantly. The demand for CPU continued to go up across all our customers. That's when we had our earnings call. We actually updated our market forecast to more than $120 billion. It's still very early when you think about agentic AI adoption. What we are seeing is with agentic AI, it's quite complex when you have millions of operations you have to execute. When you have a lot of agent concurrently working on different tasks, you actually need a very high core count CPUs. That market expansion, we saw it earlier, but we never estimate it's going to be so fast. that market expansion we saw it earlier but we never estimate it's going to be so fast The pace and the speed of agentic AI adoption has been tremendous. the pace and the speed of agentic ai adoption has been tremendous When you look at the January, February, March, all the enterprise adoption started. when you look at the january february march all the enterprise adoption started You see Anthropic, OpenAI, their revenue has been going up very significantly. you see anthropic openai their revenue has been going up very significantly The demand for CPU continued to go up across all our customers. the demand for cpu continued to go up across all our customers That's when we had our earnings call. that's when we had our earnings call We actually updated our market forecast to more than $120 billion. we actually updated our market forecast to more than $120 billion It's still very early when you think about agentic AI adoption. it's still very early when you think about agentic ai adoption What we are seeing is with agentic AI, it's quite complex when you have millions of operations you have to execute. what we are seeing is with agentic ai it's quite complex when you have millions of operations you have to execute When you have a lot of agent concurrently working on different tasks, you actually need a very high core count CPUs. when you have a lot of agent concurrently working on different tasks you actually need a very high core count cpus We don't know what other companies, how they think about the TAM and the SAM. We go through the bottom-up and the top-down approach. We know at this particular point, we feel pretty good about more than $120 billion, but it's evolving so fast, and the more complex you should expect for the agentic AI portion of the market, because you have so many diverse workloads, so complex, the ASP will continue to increase because increasingly you need a very high core count CPUs, high performance CPUs. Matt, I don't know if you have other you can add here. We don't know what other companies, how they think about the TAM and the SAM. we don't know what other companies how they think about the tam and the sam We go through the bottom-up and the top-down approach. we go through the bottom-up and the top-down approach We know at this particular point, we feel pretty good about more than $120 billion, but it's evolving so fast, and the more complex you should expect for the agentic AI portion of the market, because you have so many diverse workloads, so complex, the ASP will continue to increase because increasingly you need a very high core count CPUs, high performance CPUs. we know at this particular point we feel pretty good about more than $120 billion but it's evolving so fast and the more complex you should expect for the agentic ai portion of the market because you have so many diverse workloads so complex the asp will continue to increase because increasingly you need a very high core count cpus high performance cpus Matt, I don't know if you have other you can add here. matt i don't know if you have other you can add here

Speaker 2: Sure. First off, thank you, Vivek, for hosting us and for you guys all coming. I think there's a couple things, Vivek, that I would maybe add to Jean's comments. One is, I think for this whole audience and us included, we were trying to figure out when was the primary driver of AI CapEx going to go from almost predominantly large model training and start the shift towards inference being a primary driver of CapEx. We're seeing that happen in real time. I think the more powerful thing for our business and maybe the more transformative thing architecturally is during that transition, you're seeing chatbot inference become agentic inference, and what agentic inference does as a workload. Sure. sure First off, thank you, Vivek, for hosting us and for you guys all coming. first off thank you vivek for hosting us and for you guys all coming I think there's a couple things, Vivek, that I would maybe add to Jean's comments. i think there's a couple things vivek that i would maybe add to jean's comments One is, I think for this whole audience and us included, we were trying to figure out when was the primary driver of AI CapEx going to go from almost predominantly large model training and start the shift towards inference being a primary driver of CapEx. one is i think for this whole audience and us included we were trying to figure out when was the primary driver of ai capex going to go from almost predominantly large model training and start the shift towards inference being a primary driver of capex We're seeing that happen in real time. we're seeing that happen in real time I think the more powerful thing for our business and maybe the more transformative thing architecturally is during that transition, you're seeing chatbot inference become agentic inference, and what agentic inference does as a workload. i think the more powerful thing for our business and maybe the more transformative thing architecturally is during that transition you're seeing chatbot inference become agentic inference and what agentic inference does as a workload You daisy chain or concatenate all these automated agentic flows together, between each inference task, there's a lot of CPU diverse work, so post-processing of data, figuring out what to tell the AI to do next, often based on the result of the prior inference. Where do I get the data from for the next step? Is it in a cloud? Is it in an ERP system, a payment system, a CRM system, wherever? Get all of that data, come back, post-process the data and from that step, and feed it into the next agentic step for the AI. That workload is quite diverse. That's what we're seeing driving the demand for Turin, now in real time. You daisy chain or concatenate all these automated agentic flows together, between each inference task, there's a lot of CPU diverse work, so post-processing of data, figuring out what to tell the AI to do next, often based on the result of the prior inference. you daisy chain or concatenate all these automated agentic flows together between each inference task there's a lot of cpu diverse work so post-processing of data figuring out what to tell the ai to do next often based on the result of the prior inference Where do I get the data from for the next step? where do i get the data from for the next step Is it in a cloud? is it in a cloud Is it in an ERP system, a payment system, a CRM system, wherever? is it in an erp system a payment system a crm system wherever Get all of that data, come back, post-process the data and from that step, and feed it into the next agentic step for the AI. get all of that data come back post-process the data and from that step and feed it into the next agentic step for the ai That workload is quite diverse. that workload is quite diverse That's what we're seeing driving the demand for Turin, now in real time. that's what we're seeing driving the demand for turin now in real time As the order book fills in for the 256 core 2 nm Venice parts, that are going to launch in a couple of months and be the primary workhorse for next year. That's why we're seeing the order book really expand. As the order book fills in for the 256 core 2 nm Venice parts, that are going to launch in a couple of months and be the primary workhorse for next year. as the order book fills in for the 256 core 2 nm venice parts that are going to launch in a couple of months and be the primary workhorse for next year That's why we're seeing the order book really expand. that's why we're seeing the order book really expand

Speaker 3: Got it. Is there a way to segment that market, whatever it happens to be? $120 billion+, in those three segments, Jean, that you described, the kind of the traditional applications, let's call it part of the AI cluster and then agentic kind of standalone CPUs? Got it. got it Is there a way to segment that market, whatever it happens to be? $120 billion+ , in those three segments, Jean, that you described, the kind of the traditional applications, let's call it part of the AI cluster and then agentic kind of standalone CPUs? is there a way to segment that market whatever it happens to be $120 billion+ in those three segments jean that you described the kind of the traditional applications let's call it part of the ai cluster and then agentic kind of standalone cpus

Speaker 1: I think a traditional segment, it's very clear. I think 2025, a different third party would say that's a $25 billion-$30 billion market opportunity. It will continue to grow. As Matt has mentioned, is you actually have an agent doing more work on the traditional database. I think a traditional segment, it's very clear. i think a traditional segment it's very clear I think 2025, a different third party would say that's a $25 billion-$30 billion market opportunity. i think 2025 a different third party would say that's a $25 billion-$30 billion market opportunity It will continue to grow. it will continue to grow As Matt has mentioned, is you actually have an agent doing more work on the traditional database. as matt has mentioned is you actually have an agent doing more work on the traditional database

Speaker 3: Right. Right. right

Speaker 1: That traditional general purpose CPU, it's probably steady growth, but not as high. The head node, which is working with the GPUs, you can see that's when people talk about the ratios, is traditionally it's one head node CPU manager, eight GPUs - That traditional general purpose CPU, it's probably steady growth, but not as high. that traditional general purpose cpu it's probably steady growth but not as high The head node, which is working with the GPUs, you can see that's when people talk about the ratios, is traditionally it's one head node CPU manager, eight GPUs - the head node which is working with the gpus you can see that's when people talk about the ratios is traditionally it's one head node cpu manager eight gpus -

Speaker 3: Nice. Nice. nice

Speaker 1: - two to, yeah. It's getting more and more. Over time, that ratio is changing. We do think the head node segment CPU will also grow very fast, much faster than the traditional general CPU. The most exciting portion of the market is actually agentic AI. Agentic AI, you actually are seeing agentic AI server rack sit in between the traditional servers and the GPUs. Those racks are handling all those different workload to really make sure all the agentic agent works. That market, we think by the whatever, the $120 billion or $200 billion market opportunity is majority of that large market. - two to, yeah. - two to yeah It's getting more and more. it's getting more and more Over time, that ratio is changing. over time that ratio is changing We do think the head node segment CPU will also grow very fast, much faster than the traditional general CPU. we do think the head node segment cpu will also grow very fast much faster than the traditional general cpu The most exciting portion of the market is actually agentic AI. the most exciting portion of the market is actually agentic ai Agentic AI, you actually are seeing agentic AI server rack sit in between the traditional servers and the GPUs. agentic ai you actually are seeing agentic ai server rack sit in between the traditional servers and the gpus Those racks are handling all those different workload to really make sure all the agentic agent works. those racks are handling all those different workload to really make sure all the agentic agent works That market, we think by the whatever, the $120 billion or $200 billion market opportunity is majority of that large market. that market we think by the whatever the $120 billion or $200 billion market opportunity is majority of that large market

Speaker 3: Over 50% of - Over 50% of - over 50% of -

Speaker 1: Over 50%. Over 50%. over 50%

Speaker 3: - that market you think is that standalone? - that market you think is that standalone? - that market you think is that standalone

Speaker 1: Agentic. Yes. Agentic. agentic Yes. yes

Speaker 2: Yeah. Yeah. yeah

Speaker 1: It's still very early right now, but if you just think about what Matt said, how complex, how dynamic those workload will be, and we do see significant productivity improvement, especially software engineering side. It's still very early right now, but if you just think about what Matt said, how complex, how dynamic those workload will be, and we do see significant productivity improvement, especially software engineering side. it's still very early right now but if you just think about what matt said how complex how dynamic those workload will be and we do see significant productivity improvement especially software engineering side

Speaker 3: Nice. Nice. nice

Speaker 1: Those are very complex workload. You really need different core counts. You need really high performance, high core counts. You also need simultaneous multi-thread, all those high performance, you can manage millions of agent potentially. Those are very complex workload. those are very complex workload You really need different core counts. you really need different core counts You need really high performance, high core counts. you need really high performance high core counts You also need simultaneous multi-thread, all those high performance, you can manage millions of agent potentially. you also need simultaneous multi-thread all those high performance you can manage millions of agent potentially

Speaker 3: Got it. The final question there is actually two last questions. One is x86 versus Arm. If let's say that agentic opportunity does capture the dominant share of whatever that TAM number happens to be, what advantage, and if I dare ask, what disadvantage does x86 have versus Arm in capturing that? Got it. got it The final question there is actually two last questions. the final question there is actually two last questions One is x86 versus Arm. one is x86 versus arm If let's say that agentic opportunity does capture the dominant share of whatever that TAM number happens to be, what advantage, and if I dare ask, what disadvantage does x86 have versus Arm in capturing that? if let's say that agentic opportunity does capture the dominant share of whatever that tam number happens to be what advantage and if i dare ask what disadvantage does x86 have versus arm in capturing that

Speaker 1: I'll start. Matt can talk about the more technical side is the way AMD has been thinking about it has never been, oh, x86 architecture or Arm architecture. It's more about how we provide the best TCO for customers. For us, the performance is number one. Secondly, what we have been trying to build is the breadth and the depth. For platform portfolio. If you look at our CPU platform, we have gone through five generation at the Turin and the Venice is next generation. We have the breadth from the core count to 8 to 16 core count addressing enterprise applications to the Venice will have 256 core count. For us, we have all different configuration design points to meet all the different compute to give them best TCOs. I don't know technically you can provide them. I'll start. i'll start Matt can talk about the more technical side is the way AMD has been thinking about it has never been, oh, x86 architecture or Arm architecture. matt can talk about the more technical side is the way amd has been thinking about it has never been oh x86 architecture or arm architecture It's more about how we provide the best TCO for customers. it's more about how we provide the best tco for customers For us, the performance is number one. for us the performance is number one Secondly, what we have been trying to build is the breadth and the depth. secondly what we have been trying to build is the breadth and the depth For platform portfolio. for platform portfolio If you look at our CPU platform, we have gone through five generation at the Turin and the Venice is next generation. if you look at our cpu platform we have gone through five generation at the turin and the venice is next generation We have the breadth from the core count to 8 to 16 core count addressing enterprise applications to the Venice will have 256 core count. we have the breadth from the core count to 8 to 16 core count addressing enterprise applications to the venice will have 256 core count For us, we have all different configuration design points to meet all the different compute to give them best TCOs. for us we have all different configuration design points to meet all the different compute to give them best tcos I don't know technically you can provide them. i don't know technically you can provide them

Speaker 2: Yeah. Vivek, I think as Jean talked about, the three different buckets, right? The traditional server workloads for enterprise, I think there is an affinity towards x86 because of the code base. Many of those servers are deployed on-prem, some of them are deployed in third-party cloud. It's basically the same workload depending on where you want to deploy. I think as these agents generate more traffic to all those traditional systems, the x86 ecosystem is well positioned there. You have the head nodes. I think that what you really need in head nodes is really high single-thread performance, really high bandwidth to, I don't know, it's not a technical term, but feed the beast of these really expensive accelerators that are in the box, right? We will do our own head nodes for Helios. Yeah. yeah Vivek, I think as Jean talked about, the three different buckets, right? vivek i think as jean talked about the three different buckets right The traditional server workloads for enterprise, I think there is an affinity towards x86 because of the code base. the traditional server workloads for enterprise i think there is an affinity towards x86 because of the code base Many of those servers are deployed on-prem, some of them are deployed in third-party cloud. many of those servers are deployed on-prem some of them are deployed in third-party cloud It's basically the same workload depending on where you want to deploy. it's basically the same workload depending on where you want to deploy I think as these agents generate more traffic to all those traditional systems, the x86 ecosystem is well positioned there. i think as these agents generate more traffic to all those traditional systems the x86 ecosystem is well positioned there You have the head nodes. you have the head nodes I think that what you really need in head nodes is really high single-thread performance, really high bandwidth to, I don't know, it's not a technical term, but feed the beast of these really expensive accelerators that are in the box, right? i think that what you really need in head nodes is really high single-thread performance really high bandwidth to i don't know it's not a technical term but feed the beast of these really expensive accelerators that are in the box right We will do our own head nodes for Helios. we will do our own head nodes for helios I think our large competitor in their NVL racks will basically do their own head nodes. There's a bunch of different XPU programs that are going to be launched by the industry. I think when we have those conversations with customers, it's much more about what is the performance of your CPU such that I get the best utilization out of this really expensive XPU that I'm buying. It's not an Intel or AMD or x86 versus Arm conversation. It's a what's the best server part. When you talk about these agentic racks, it's really about how many concurrent agents can you run per rack or per megawatt of compute. As Jean mentioned, really high thread and core count products. I think our roadmap is differentiated there. I think our large competitor in their NVL racks will basically do their own head nodes. i think our large competitor in their nvl racks will basically do their own head nodes There's a bunch of different XPU programs that are going to be launched by the industry. there's a bunch of different xpu programs that are going to be launched by the industry I think when we have those conversations with customers, it's much more about what is the performance of your CPU such that I get the best utilization out of this really expensive XPU that I'm buying. i think when we have those conversations with customers it's much more about what is the performance of your cpu such that i get the best utilization out of this really expensive xpu that i'm buying It's not an Intel or AMD or x86 versus Arm conversation. it's not an intel or amd or x86 versus arm conversation It's a what's the best server part. it's a what's the best server part When you talk about these agentic racks, it's really about how many concurrent agents can you run per rack or per megawatt of compute. when you talk about these agentic racks it's really about how many concurrent agents can you run per rack or per megawatt of compute As Jean mentioned, really high thread and core count products. as jean mentioned really high thread and core count products I think our roadmap is differentiated there. i think our roadmap is differentiated there Another thing that doesn't get mentioned, but I think is important, is if you think about the dollars of CapEx that are going to come into the server market, the mission-critical nature of these systems, both head node and agentic racks. Some of the RAS features and the security features of the x86 ecosystem, it's not an x86 architecture thing, it's just we and our x86 competitor have been put through the paces by every hyperscaler and most enterprises on all of these security features for a decade. You might imagine, you get into these automated flows where those agents have access to mission-critical data, and you want to really have servers that have robust security features there. We feel we get brought into essentially every RFQ that's in the industry for these servers, and I think we're positioned to win a very high percentage of them. Another thing that doesn't get mentioned, but I think is important, is if you think about the dollars of CapEx that are going to come into the server market, the mission-critical nature of these systems, both head node and agentic racks. another thing that doesn't get mentioned but i think is important is if you think about the dollars of capex that are going to come into the server market the mission-critical nature of these systems both head node and agentic racks Some of the RAS features and the security features of the x86 ecosystem, it's not an x86 architecture thing, it's just we and our x86 competitor have been put through the paces by every hyperscaler and most enterprises on all of these security features for a decade. some of the ras features and the security features of the x86 ecosystem it's not an x86 architecture thing it's just we and our x86 competitor have been put through the paces by every hyperscaler and most enterprises on all of these security features for a decade You might imagine, you get into these automated flows where those agents have access to mission-critical data, and you want to really have servers that have robust security features there. you might imagine you get into these automated flows where those agents have access to mission-critical data and you want to really have servers that have robust security features there We feel we get brought into essentially every RFQ that's in the industry for these servers, and I think we're positioned to win a very high percentage of them. we feel we get brought into essentially every rfq that's in the industry for these servers and i think we're positioned to win a very high percentage of them

Speaker 3: Got it. Now, your x86 competitor has their own fab. They are planning to add a lot more capacity on the server CPU side. I guess this year every CPU is being sold, right? Do you think that there is a scope for share shifts given that your competitor has incremental capacity to devote to this, or you think you will be adequately served by your foundry ecosystem? Got it. got it Now, your x86 competitor has their own fab. now your x86 competitor has their own fab They are planning to add a lot more capacity on the server CPU side. they are planning to add a lot more capacity on the server cpu side I guess this year every CPU is being sold, right? i guess this year every cpu is being sold right Do you think that there is a scope for share shifts given that your competitor has incremental capacity to devote to this, or you think you will be adequately served by your foundry ecosystem? do you think that there is a scope for share shifts given that your competitor has incremental capacity to devote to this or you think you will be adequately served by your foundry ecosystem

Speaker 1: Yeah. We have been very pleased with our share gain trajectory. I think the last Q1, which is announced, we actually from a value share perspective, we got to like a 46%. Yeah. yeah We have been very pleased with our share gain trajectory. we have been very pleased with our share gain trajectory I think the last Q1, which is announced, we actually from a value share perspective, we got to like a 46%. i think the last q1 which is announced we actually from a value share perspective we got to like a 46%

Speaker 3: Nice. Nice. nice

Speaker 1: When you think about it, our success is really tied together with TSMC. From day one, we have been using TSMC. We work together with our chiplet design, with the packaging technology to come up with the best performance server platforms. That continue to benefit us going forward. We have been planning for this ramp since last year. If you look at the Q1 performance and the Q2 guidance, when we guide year-over-year 70% increase. A lot of you actually know is right, the way we first started six or nine months ago. It is the early planning, how we work with the TSMC to make sure we get support to continue to drive the ramp of the supply. The demand is tremendous. Supply is still tight. When you think about it, our success is really tied together with TSMC. when you think about it our success is really tied together with tsmc From day one, we have been using TSMC. from day one we have been using tsmc We work together with our chiplet design, with the packaging technology to come up with the best performance server platforms. we work together with our chiplet design with the packaging technology to come up with the best performance server platforms That continue to benefit us going forward. that continue to benefit us going forward We have been planning for this ramp since last year. we have been planning for this ramp since last year If you look at the Q1 performance and the Q2 guidance, when we guide year-over-year 70% increase. if you look at the q1 performance and the q2 guidance when we guide year-over-year 70% increase A lot of you actually know is right, the way we first started six or nine months ago. a lot of you actually know is right the way we first started six or nine months ago It is the early planning, how we work with the TSMC to make sure we get support to continue to drive the ramp of the supply. it is the early planning how we work with the tsmc to make sure we get support to continue to drive the ramp of the supply The demand is tremendous. the demand is tremendous Supply is still tight. supply is still tight I think for us, for this year, next year, every supply we can get, we can provide to our customers. We're going to continue to do that. It's the planning beyond the 2027 right now. For the longer term, I think we'll continue to work with TSMC and prioritize data center as how we drive the business growth going forward. I think for us, for this year, next year, every supply we can get, we can provide to our customers. i think for us for this year next year every supply we can get we can provide to our customers We're going to continue to do that. we're going to continue to do that It's the planning beyond the 2027 right now. it's the planning beyond the 2027 right now For the longer term, I think we'll continue to work with TSMC and prioritize data center as how we drive the business growth going forward. for the longer term i think we'll continue to work with tsmc and prioritize data center as how we drive the business growth going forward

Speaker 3: Got it. Got it. got it

Speaker 2: Vivek, one little thing that I would add to what Jean mentioned there is when we have conversations with the investor community on this particular topic, I think one thing that I've noticed is supply is tight, right? 3 nm tight. There's a lot of other areas that are tight, and I think we're very well-positioned with the relationships Lisa has personally with the executives in that space to get maybe more than our fair share of the incremental. It is tight. I think what I've noticed is maybe the investor scoping of what Lisa and the team initially asked for in 2026 and 2027 maybe didn't imagine how much growth we had already planned for in supply. Vivek, one little thing that I would add to what Jean mentioned there is when we have conversations with the investor community on this particular topic, I think one thing that I've noticed is supply is tight, right? 3 nm tight. vivek one little thing that i would add to what jean mentioned there is when we have conversations with the investor community on this particular topic i think one thing that i've noticed is supply is tight right 3 nm tight There's a lot of other areas that are tight, and I think we're very well-positioned with the relationships Lisa has personally with the executives in that space to get maybe more than our fair share of the incremental. there's a lot of other areas that are tight and i think we're very well-positioned with the relationships lisa has personally with the executives in that space to get maybe more than our fair share of the incremental It is tight. it is tight I think what I've noticed is maybe the investor scoping of what Lisa and the team initially asked for in 2026 and 2027 maybe didn't imagine how much growth we had already planned for in supply. i think what i've noticed is maybe the investor scoping of what lisa and the team initially asked for in 2026 and 2027 maybe didn't imagine how much growth we had already planned for in supply Things are tight, but we've already asked for and already been sort of, quote-unquote, "granted" from TSMC, I think we're having the right conversation there, but maybe not at the right starting point from what we were sort of allocated initially. Things are tight, but we've already asked for and already been sort of, quote-unquote, "granted" from TSMC, I think we're having the right conversation there, but maybe not at the right starting point from what we were sort of allocated initially. things are tight but we've already asked for and already been sort of quote-unquote "granted" from tsmc i think we're having the right conversation there but maybe not at the right starting point from what we were sort of allocated initially

Speaker 3: Got it. Final one on the CPU side on price. You mentioned 50%+ growth in Q1 and then 70%+. I think there was a sense that one could sustain this - Got it. got it Final one on the CPU side on price. final one on the cpu side on price You mentioned 50%+ growth in Q1 and then 70%+ . you mentioned 50%+ growth in q1 and then 70%+ I think there was a sense that one could sustain this - i think there was a sense that one could sustain this -

Speaker 2: Yeah. Yeah. yeah

Speaker 3: - for the rest of the year. How much of this, Jean, is pricing versus unit growth? Let me just add quickly part B to that. As that agentic, that third segment, becomes a bigger portion, do you think that can command much higher ASPs? When I talk with the Arm camp, they're talking about ASPs in the $3,000, $4,000, $5,000 range, which is well above the ASPs that AMD is able to get. Just unit versus ASP and just the trend of ASP. - for the rest of the year. - for the rest of the year How much of this, Jean, is pricing versus unit growth? how much of this jean is pricing versus unit growth Let me just add quickly part B to that. As that agentic, that third segment, becomes a bigger portion, do you think that can command much higher ASPs? let me just add quickly part b to that. as that agentic that third segment becomes a bigger portion do you think that can command much higher asps When I talk with the Arm camp, they're talking about ASPs in the $3,000, $4,000, $5,000 range, which is well above the ASPs that AMD is able to get. when i talk with the arm camp they're talking about asps in the $3,000 $4,000 $5,000 range which is well above the asps that amd is able to get Just unit versus ASP and just the trend of ASP. just unit versus asp and just the trend of asp

Speaker 1: Yeah. I think as you mentioned, both the more than 50% growth in Q1 and that we guide the Q2 70% year-over-year, more than 70% year-over-year increase, two-third of that actually are unit growth. Majority of our revenue growth actually come from unit expansion versus ASP increase. Your point on the ASP going forward, especially with the agentic AI, is a great point. What we're seeing is we're at the early inning of agentic AI adoption. You can already see with the software engineering part of agentic work, autonomous agentic workflow, you really need a very high performance, high core counts, CPUs. Yeah. yeah I think as you mentioned, both the more than 50% growth in Q1 and that we guide the Q2 70% year-over-year, more than 70% year-over-year increase, two-third of that actually are unit growth. i think as you mentioned both the more than 50% growth in q1 and that we guide the q2 70% year-over-year more than 70% year-over-year increase two-third of that actually are unit growth Majority of our revenue growth actually come from unit expansion versus ASP increase. majority of our revenue growth actually come from unit expansion versus asp increase Your point on the ASP going forward, especially with the agentic AI, is a great point. your point on the asp going forward especially with the agentic ai is a great point What we're seeing is we're at the early inning of agentic AI adoption. what we're seeing is we're at the early inning of agentic ai adoption You can already see with the software engineering part of agentic work, autonomous agentic workflow, you really need a very high performance, high core counts, CPUs. you can already see with the software engineering part of agentic work autonomous agentic workflow you really need a very high performance high core counts cpus

Speaker 3: Right. Right. right

Speaker 1: That tend to have a higher ASP. I think a generation over generation, when we go to 2 nm with Venice and the future generation to handle complex workload to provide the performance, you tend to see the CPU price continue to increase. The other thing I would just say is because we have a really large platform, the CPU ASP is really determined by the mix - That tend to have a higher ASP. that tend to have a higher asp I think a generation over generation, when we go to 2 nm with Venice and the future generation to handle complex workload to provide the performance, you tend to see the CPU price continue to increase. i think a generation over generation when we go to 2 nm with venice and the future generation to handle complex workload to provide the performance you tend to see the cpu price continue to increase The other thing I would just say is because we have a really large platform, the CPU ASP is really determined by the mix - the other thing i would just say is because we have a really large platform the cpu asp is really determined by the mix -

Speaker 3: Right. Right. right

Speaker 1: - by the different configurations for different workload. For enterprise, the per core ASP is very high, but because the core counts are relatively low, so ASP is relatively low, but the gross margin is great. It's very nuanced from our business because how big our platform is and how complex all different design point we're helping our customers. - by the different configurations for different workload. - by the different configurations for different workload For enterprise, the per core ASP is very high, but because the core counts are relatively low, so ASP is relatively low, but the gross margin is great. for enterprise the per core asp is very high but because the core counts are relatively low so asp is relatively low but the gross margin is great It's very nuanced from our business because how big our platform is and how complex all different design point we're helping our customers. it's very nuanced from our business because how big our platform is and how complex all different design point we're helping our customers

Speaker 3: Got it. You spent two thirds of the time on CPU, let's talk about GPU. Got it. got it You spent two thirds of the time on CPU, let's talk about GPU. you spent two thirds of the time on cpu let's talk about gpu

Speaker 1: It's great. It's great. it's great

Speaker 3: GPU now. You are at the start of launch of a major rack-scale system to your portfolio. Maybe give us a status update of how you're feeling. When your competitor launched their first large scale up product, it took a while for the industry to get ready, liquid cooling and all those things to get ready. How do you think about your progress so far? How excited are you about the second half? GPU now. gpu now You are at the start of launch of a major rack-scale system to your portfolio. you are at the start of launch of a major rack-scale system to your portfolio Maybe give us a status update of how you're feeling. maybe give us a status update of how you're feeling When your competitor launched their first large scale up product, it took a while for the industry to get ready, liquid cooling and all those things to get ready. when your competitor launched their first large scale up product it took a while for the industry to get ready liquid cooling and all those things to get ready How do you think about your progress so far? how do you think about your progress so far How excited are you about the second half? how excited are you about the second half

Speaker 1: Yeah, we are on track. We sampled MI450. We expect to be launching in second half or Q3 starting point and the Q4. As you know, we have been preparing for this launch for a long time. We acquired ZT Systems in 2024. The whole reason we're acquiring ZT Systems is to help us to add capabilities in system-level design. We also have worked on the networking side. The whole team has been working to prepare this launch. We know it's very hard. It's very complex, very hard, but from execution perspective, our team has been on track, meticulous prepare for all different aspect launch. Yeah, we are on track. yeah we are on track We sampled MI450. we sampled mi450 We expect to be launching in second half or Q3 starting point and the Q4. we expect to be launching in second half or q3 starting point and the q4 As you know, we have been preparing for this launch for a long time. as you know we have been preparing for this launch for a long time We acquired ZT Systems in 2024. we acquired zt systems in 2024 The whole reason we're acquiring ZT Systems is to help us to add capabilities in system-level design. the whole reason we're acquiring zt systems is to help us to add capabilities in system-level design We also have worked on the networking side. we also have worked on the networking side The whole team has been working to prepare this launch. the whole team has been working to prepare this launch We know it's very hard. we know it's very hard It's very complex, very hard, but from execution perspective, our team has been on track, meticulous prepare for all different aspect launch. it's very complex very hard but from execution perspective our team has been on track meticulous prepare for all different aspect launch Not only technology software side, but more importantly, the supply chain side, how you make sure the design redundancy with all the components, but also supplies even very small components, you have to make sure you have supplies. That has been the whole team's execution track record. We feel pretty good right now. Not only technology software side, but more importantly, the supply chain side, how you make sure the design redundancy with all the components, but also supplies even very small components, you have to make sure you have supplies. not only technology software side but more importantly the supply chain side how you make sure the design redundancy with all the components but also supplies even very small components you have to make sure you have supplies That has been the whole team's execution track record. that has been the whole team's execution track record We feel pretty good right now. we feel pretty good right now

Speaker 2: Vivek, I think we're anticipating, as Jean said, we already have sampled and we have a number of customers that have full Helios racks in their own data center running their full production workload now and testing and doing all those things that you would expect us to be doing at this point in time. The customers learned a lot, we learned a lot from some folks that blazed the trail, as you mentioned. We're going to really focus on a concentrated set of ODMs as we launch for the first number of months until we get up to scale, not like a few systems, but billions of dollars of scale. We'll be expanding the ODM ecosystem pretty significantly as we move through next year and get towards the MI500 series and whatnot. We feel good about where we are. Vivek, I think we're anticipating, as Jean said, we already have sampled and we have a number of customers that have full Helios racks in their own data center running their full production workload now and testing and doing all those things that you would expect us to be doing at this point in time. vivek i think we're anticipating as jean said we already have sampled and we have a number of customers that have full helios racks in their own data center running their full production workload now and testing and doing all those things that you would expect us to be doing at this point in time The customers learned a lot, we learned a lot from some folks that blazed the trail, as you mentioned. the customers learned a lot we learned a lot from some folks that blazed the trail as you mentioned We're going to really focus on a concentrated set of ODMs as we launch for the first number of months until we get up to scale, not like a few systems, but billions of dollars of scale. we're going to really focus on a concentrated set of odms as we launch for the first number of months until we get up to scale not like a few systems but billions of dollars of scale We'll be expanding the ODM ecosystem pretty significantly as we move through next year and get towards the MI500 series and whatnot. we'll be expanding the odm ecosystem pretty significantly as we move through next year and get towards the mi500 series and whatnot We feel good about where we are. we feel good about where we are The initial work that the customer base is doing in their own labs with sampled systems has gone very well. Now it's about We'll see a pretty significant jump in the fourth quarter of revenue, another fairly significant jump in Q1, and then we'll go from there. It is wood, so touch wood. Everything's going exactly as we'd hoped it goes. When you're doing a thing like this, day to day, there's always something. Our job's to keep the duck calm on the top as we kick on the bottom and get ready to ramp this thing starting in Q3. The initial work that the customer base is doing in their own labs with sampled systems has gone very well. the initial work that the customer base is doing in their own labs with sampled systems has gone very well Now it's about We'll see a pretty significant jump in the fourth quarter of revenue, another fairly significant jump in Q1, and then we'll go from there. now it's about we'll see a pretty significant jump in the fourth quarter of revenue another fairly significant jump in q1 and then we'll go from there It is wood, so touch wood. it is wood so touch wood Everything's going exactly as we'd hoped it goes. everything's going exactly as we'd hoped it goes When you're doing a thing like this, day to day, there's always something. when you're doing a thing like this day to day there's always something Our job's to keep the duck calm on the top as we kick on the bottom and get ready to ramp this thing starting in Q3. our job's to keep the duck calm on the top as we kick on the bottom and get ready to ramp this thing starting in q3

Speaker 3: Got it. When you've been asked, Jean, or Lisa has been asked, you have always mentioned the opportunity for multi-gigawatt scale deployments. I think we have heard about two of them, OpenAI and Meta. You think there is still the scope for us to be pleasantly surprised with more such announcements even for 2027? You think that just given how tight the industry is, that more incremental announcements could be further spaced out than that? Got it. got it When you've been asked, Jean, or Lisa has been asked, you have always mentioned the opportunity for multi-gigawatt scale deployments. when you've been asked jean or lisa has been asked you have always mentioned the opportunity for multi-gigawatt scale deployments I think we have heard about two of them, OpenAI and Meta. i think we have heard about two of them openai and meta You think there is still the scope for us to be pleasantly surprised with more such announcements even for 2027? you think there is still the scope for us to be pleasantly surprised with more such announcements even for 2027 You think that just given how tight the industry is, that more incremental announcements could be further spaced out than that? you think that just given how tight the industry is that more incremental announcements could be further spaced out than that

Speaker 1: Yeah. We talked about we're really pleased with our partnership, very long-term partnership with both OpenAI and Meta. Last year, we established those relationship, and we actually see the forecast from those customers are actually above our original plan for 2027. I think one of the key milestone we're really focusing on this year is to continue to expand and deepening the customer adoption of MI450. We definitely expect there will be other multi-gigawatt customers. We have both, across the board, hyperscale customers, the model companies, and even AI native companies who we have been engaging, working with. I think it's important for us to continue to expand this customer base and with the largest scale deployment. Yeah. yeah We talked about we're really pleased with our partnership, very long-term partnership with both OpenAI and Meta. we talked about we're really pleased with our partnership very long-term partnership with both openai and meta Last year, we established those relationship, and we actually see the forecast from those customers are actually above our original plan for 2027. last year we established those relationship and we actually see the forecast from those customers are actually above our original plan for 2027 I think one of the key milestone we're really focusing on this year is to continue to expand and deepening the customer adoption of MI450. i think one of the key milestone we're really focusing on this year is to continue to expand and deepening the customer adoption of mi450 We definitely expect there will be other multi-gigawatt customers. we definitely expect there will be other multi-gigawatt customers We have both, across the board, hyperscale customers, the model companies, and even AI native companies who we have been engaging, working with. we have both across the board hyperscale customers the model companies and even ai native companies who we have been engaging working with I think it's important for us to continue to expand this customer base and with the largest scale deployment. i think it's important for us to continue to expand this customer base and with the largest scale deployment

Speaker 3: Got it. Makes sense. How do you expect the two customers you have announced to allocate market share? Because they have a number of supplier options. I know you have a somewhat unique engagement with them through the granting of warrants and whatnot, but just if we set that aside, how is your visibility into getting, I think the two gigawatt each from those customers for next year? Got it. got it Makes sense. makes sense How do you expect the two customers you have announced to allocate market share? how do you expect the two customers you have announced to allocate market share Because they have a number of supplier options. because they have a number of supplier options I know you have a somewhat unique engagement with them through the granting of warrants and whatnot, but just if we set that aside, how is your visibility into getting, I think the two gigawatt each from those customers for next year? i know you have a somewhat unique engagement with them through the granting of warrants and whatnot but just if we set that aside how is your visibility into getting i think the two gigawatt each from those customers for next year

Speaker 1: Vivek, as you know, the lead time for MI450 is quite long. We really have to start to plan and have the orders from customers for next year for the second half launch and the next year. I think the supply chain overall is tight next year. Since we have been preparing to support our customers, we do want to ensure we get the supplies for what they need. Those are two our most important anchor customers. Of course, we also have a warrant, which we're going to be able to share the upside if we get more revenue. Our stock price is higher, they will share upside too. That is absolutely the common interest for both companies who want to drive the adoption of MI450. We feel pretty good with both of them. We think we can get the right supplies to support them. Vivek, as you know, the lead time for MI450 is quite long. vivek as you know the lead time for mi450 is quite long We really have to start to plan and have the orders from customers for next year for the second half launch and the next year. we really have to start to plan and have the orders from customers for next year for the second half launch and the next year I think the supply chain overall is tight next year. i think the supply chain overall is tight next year Since we have been preparing to support our customers, we do want to ensure we get the supplies for what they need. since we have been preparing to support our customers we do want to ensure we get the supplies for what they need Those are two our most important anchor customers. those are two our most important anchor customers Of course, we also have a warrant, which we're going to be able to share the upside if we get more revenue. of course we also have a warrant which we're going to be able to share the upside if we get more revenue Our stock price is higher, they will share upside too. our stock price is higher they will share upside too That is absolutely the common interest for both companies who want to drive the adoption of MI450. that is absolutely the common interest for both companies who want to drive the adoption of mi450 We feel pretty good with both of them. we feel pretty good with both of them We think we can get the right supplies to support them. we think we can get the right supplies to support them

Speaker 3: Got it. Do you think having someone like TSMC is almost adding a level of discipline in the ecosystem? It's just harder to double count if everyone has to go to the same source. I know it sounds obvious, but do you see that also? Got it. got it Do you think having someone like TSMC is almost adding a level of discipline in the ecosystem? do you think having someone like tsmc is almost adding a level of discipline in the ecosystem It's just harder to double count if everyone has to go to the same source. it's just harder to double count if everyone has to go to the same source I know it sounds obvious, but do you see that also? i know it sounds obvious but do you see that also

Speaker 2: They're quite disciplined. I think they'll continue to be. These are sensitive topics for us to talk about in these forums, but we really, really, really appreciate the support that TSMC giving us as a partner, and the way that they're working with us to bring more capacity online to support not just the Helios and 450 growth, but the significant server TAM expansion that we're seeing and planning for not just for this year, but for all of next year and conversations going into 2028 on wafer capacity to support demand. They're quite disciplined. they're quite disciplined I think they'll continue to be. i think they'll continue to be These are sensitive topics for us to talk about in these forums, but we really, really, really appreciate the support that TSMC giving us as a partner, and the way that they're working with us to bring more capacity online to support not just the Helios and 450 growth, but the significant server TAM expansion that we're seeing and planning for not just for this year, but for all of next year and conversations going into 2028 on wafer capacity to support demand. these are sensitive topics for us to talk about in these forums but we really really really appreciate the support that tsmc giving us as a partner and the way that they're working with us to bring more capacity online to support not just the helios and 450 growth but the significant server tam expansion that we're seeing and planning for not just for this year but for all of next year and conversations going into 2028 on wafer capacity to support demand We really appreciate the work that they're doing on our behalf, and I think they're being, as they've always been, disciplined in the way that they bring capacity online and do their diligence two or three levels deep in the customer base to understand where the organic demand is coming from to support the capacity that they're bringing on. We really appreciate the work that they're doing on our behalf, and I think they're being, as they've always been, disciplined in the way that they bring capacity online and do their diligence two or three levels deep in the customer base to understand where the organic demand is coming from to support the capacity that they're bringing on. we really appreciate the work that they're doing on our behalf and i think they're being as they've always been disciplined in the way that they bring capacity online and do their diligence two or three levels deep in the customer base to understand where the organic demand is coming from to support the capacity that they're bringing on

Speaker 3: Makes sense. The final one, Jean, is on just memory cost inflation. I guess, when you wake up and see DRAM, I am sure DRAM price is not what you see when you wake up, but when you see that cost inflation every day, is that a good thing because it gives you more pricing power, or is that a bad thing because now you have to absorb that cost and hunt for more supply? What are kind of the pros and cons of that memory cost inflation for you? Makes sense. makes sense The final one, Jean, is on just memory cost inflation. the final one jean is on just memory cost inflation I guess, when you wake up and see DRAM, I am sure DRAM price is not what you see when you wake up, but when you see that cost inflation every day, is that a good thing because it gives you more pricing power, or is that a bad thing because now you have to absorb that cost and hunt for more supply? i guess when you wake up and see dram i am sure dram price is not what you see when you wake up but when you see that cost inflation every day is that a good thing because it gives you more pricing power or is that a bad thing because now you have to absorb that cost and hunt for more supply What are kind of the pros and cons of that memory cost inflation for you? what are kind of the pros and cons of that memory cost inflation for you

Speaker 1: It's really a great question. First for us, we really want to make sure if there's a longer-term agreement with our suppliers, we can secure the capacity. That is number one thing when you have a larger business at our scale, you have to plan ahead. Secondly, we think more strategically, if there's memory cost increase, we do need to figure out a way to share with our customers to absorb that increase together. I think the most important thing for us is we actually don't want to use the shortage of controlled memory to just increase the price. We want to make sure customer help us to absorb the cost increase, we are more strategic with the long-term relationship. I think from industry perspective, the memory cost increase at this kind of level, we have never seen. The price is really high. It's really a great question. it's really a great question First for us, we really want to make sure if there's a longer-term agreement with our suppliers, we can secure the capacity. first for us we really want to make sure if there's a longer-term agreement with our suppliers we can secure the capacity That is number one thing when you have a larger business at our scale, you have to plan ahead. that is number one thing when you have a larger business at our scale you have to plan ahead Secondly, we think more strategically, if there's memory cost increase, we do need to figure out a way to share with our customers to absorb that increase together. secondly we think more strategically if there's memory cost increase we do need to figure out a way to share with our customers to absorb that increase together I think the most important thing for us is we actually don't want to use the shortage of controlled memory to just increase the price. i think the most important thing for us is we actually don't want to use the shortage of controlled memory to just increase the price We want to make sure customer help us to absorb the cost increase, we are more strategic with the long-term relationship. we want to make sure customer help us to absorb the cost increase we are more strategic with the long-term relationship I think from industry perspective, the memory cost increase at this kind of level, we have never seen. i think from industry perspective the memory cost increase at this kind of level we have never seen The price is really high. the price is really high I think it does increase the CapEx spending for end customers where they spend the CapEx. I think in the longer term, just like all the industry, you always have a new capacity add up. You also have a new innovation on how you utilize memory more efficiently. It will figure out itself in the longer term. That's what we believe. I think it does increase the CapEx spending for end customers where they spend the CapEx. i think it does increase the capex spending for end customers where they spend the capex I think in the longer term, just like all the industry, you always have a new capacity add up. i think in the longer term just like all the industry you always have a new capacity add up You also have a new innovation on how you utilize memory more efficiently. you also have a new innovation on how you utilize memory more efficiently It will figure out itself in the longer term. it will figure out itself in the longer term That's what we believe. that's what we believe

Speaker 3: Got it. Do you have long-term agreements in place for the majority of your memory needs, or do you think that this is a cyclical high, and maybe there will be more normalization of that input cost for you? Got it. got it Do you have long-term agreements in place for the majority of your memory needs, or do you think that this is a cyclical high, and maybe there will be more normalization of that input cost for you? do you have long-term agreements in place for the majority of your memory needs or do you think that this is a cyclical high and maybe there will be more normalization of that input cost for you

Speaker 1: From planning perspective, because our manufacturing cycle is so long, if we plan for the MI450 ramp this year, we had to start to talk to our suppliers last year or year before. Similar situation is right now, not only with the very tight supply environment, not only we're planning 2027, we're planning 2028. For us to continue to make sure we can support our ambitious plan to grow the company, we have to plan ahead. From planning perspective, because our manufacturing cycle is so long, if we plan for the MI450 ramp this year, we had to start to talk to our suppliers last year or year before. from planning perspective because our manufacturing cycle is so long if we plan for the mi450 ramp this year we had to start to talk to our suppliers last year or year before Similar situation is right now, not only with the very tight supply environment, not only we're planning 2027, we're planning 2028. similar situation is right now not only with the very tight supply environment not only we're planning 2027 we're planning 2028 For us to continue to make sure we can support our ambitious plan to grow the company, we have to plan ahead. for us to continue to make sure we can support our ambitious plan to grow the company we have to plan ahead

Speaker 2: Yeah. Vivek, in the server market and in the PC market, the vast majority of the DRAM is actually acquired by our OEM and ODM partners and the hyperscalers themselves. That DRAM doesn't run through our P&L, and we work really, really hard with all of the customer base to try to make sure that we're matching the memory supply that they're able to get and the CPU supply or the GPU supply that we're able to provide together to where there's not mismatching of supply in the industry or hoarding of anything. We've been working really hard on that. There's certainly increased input costs to end devices in certain parts of the consumer and gaming market where there have been some anticipated market impacts, and I think we've given you guys our current view of that and the guidance that we've talked about. Yeah. yeah Vivek, in the server market and in the PC market, the vast majority of the DRAM is actually acquired by our OEM and ODM partners and the hyperscalers themselves. vivek in the server market and in the pc market the vast majority of the dram is actually acquired by our oem and odm partners and the hyperscalers themselves That DRAM doesn't run through our P&L, and we work really, really hard with all of the customer base to try to make sure that we're matching the memory supply that they're able to get and the CPU supply or the GPU supply that we're able to provide together to where there's not mis matching of supply in the industry or hoarding of anything. that dram doesn't run through our p&l and we work really really hard with all of the customer base to try to make sure that we're matching the memory supply that they're able to get and the cpu supply or the gpu supply that we're able to provide together to where there's not mis matching of supply in the industry or hoarding of anything We've been working really hard on that. we've been working really hard on that There's certainly increased input costs to end devices in certain parts of the consumer and gaming market where there have been some anticipated market impacts, and I think we've given you guys our current view of that and the guidance that we've talked about. there's certainly increased input costs to end devices in certain parts of the consumer and gaming market where there have been some anticipated market impacts and i think we've given you guys our current view of that and the guidance that we've talked about We're working really hard on behalf of the whole customer base to try to make sure that the memory pricing and the matching of components is relatively seamless in a time like this. We're working really hard on behalf of the whole customer base to try to make sure that the memory pricing and the matching of components is relatively seamless in a time like this. we're working really hard on behalf of the whole customer base to try to make sure that the memory pricing and the matching of components is relatively seamless in a time like this

Speaker 3: Got you. Just final question before we close. When you look across your customer base, Jean, how far is the visibility that they're willing to provide to you now, and when you get that, what are you doing differently in your business operationally to kind of deliver to those? Got you. got you Just final question before we close. just final question before we close When you look across your customer base, Jean, how far is the visibility that they're willing to provide to you now, and when you get that, what are you doing differently in your business operationally to kind of deliver to those? when you look across your customer base jean how far is the visibility that they're willing to provide to you now and when you get that what are you doing differently in your business operationally to kind of deliver to those

Speaker 1: Yeah. I think because the supply chain environment is really tight, customer really plan ahead. We have a very good visibility into 2027 and beyond. I think our customers and ourselves, especially when you think about the data center investment, very large CapEx planning, very long planning cycle, they actually plan way ahead. 2027, we feel really good about both the visibility of demand side and the visibility of supply side. Of course, 2028 and beyond, that's everybody's working on. Yeah. yeah I think because the supply chain environment is really tight, customer really plan ahead. i think because the supply chain environment is really tight customer really plan ahead We have a very good visibility into 2027 and beyond. we have a very good visibility into 2027 and beyond I think our customers and ourselves, especially when you think about the data center investment, very large CapEx planning, very long planning cycle, they actually plan way ahead. 2027, we feel really good about both the visibility of demand side and the visibility of supply side. i think our customers and ourselves especially when you think about the data center investment very large capex planning very long planning cycle they actually plan way ahead 2027 we feel really good about both the visibility of demand side and the visibility of supply side Of course, 2028 and beyond, that's everybody's working on. of course 2028 and beyond that's everybody's working on

Speaker 3: Okay. On that positive note, thank you so much, Jean. Okay. okay On that positive note, thank you so much, Jean. on that positive note thank you so much jean

Speaker 1: Thank you. Thank you. thank you

Speaker 3: I appreciate your time. I appreciate your time. i appreciate your time