Good afternoon, everybody. Welcome to the Goldman Sachs Communacopia + Technology Conference. I'm Jim Schneider, the semiconductor analyst here at Goldman Sachs. It's my pleasure to welcome Intel's VP of Corporate Planning and Investor Relations, John Pitzer. Welcome, John. Thanks, Jim. Thanks for being here. Yep. Before we begin, please note that today's discussion may contain forward-looking statements that are subject to various risks and uncertainties and may reference non-GAAP financial measures. Please refer to Intel's most recent earnings release and the annual report on Form 10-K and other filings with the SEC for more information on the risk factors that could cause actual results to differ materially and additional information on our non-GAAP financial measures, including reconciliations where appropriate to the corresponding GAAP financial measures. Mouthful, but anyway, thanks for being here again. Appreciate it. Thank you. I want to bring it sort of up to a high-level question for a second. Your CEO, Lip-Bu Tan, has been head of Intel for about six months now. I think you've been pretty clear about the types of changes Lip-Bu wants to bring from a sharpened focus to more internal accountability to a more outcome-based view on CapEx. Can you maybe talk a little bit about the most important changes that he's brought to the company? Maybe what are some of the most important changes internally at the company that investors might have trouble seeing? Yeah, it's a good question, Jim. I think a lot of it revolves around culture and the changing culture that Lip Bu's trying to affect at the company. I think a lot of that occurred in Q2 when we did some of the restructuring. If you remember, we did restructuring a year ago Q2. I would argue that that was a cost-cutting exercise only. It didn't really change the way that we operated the business or structured the business. I would actually argue what we did in Q2 of this year was much more about changing organizational structure and operation. That led to, I think, reduction in cost, pretty significant reduction in cost. The biggest thing that Lip-Bu really wanted to tackle was the bureaucracy inside the organization, which he thought was driving to slow and poor decisions. I think one of the things we've talked about when he joined, there were about 11 layers of management. With the actions we took in Q2, we effectively cut that in half. He's trying to drive, I think, a flatter organization with more accountability. The other thing that he did was actually mandate a return to office, which we implemented a week ago, Tuesday. About a week into it. I have yet to do that because I was in New York last week at investor conferences. I'm here this week, but I'll be down in Santa Clara tomorrow. I think it's really important over the next 4-5 months, now that we've got the team in place and we've got people coming back to the office, to really get buy-in on the strategy and see the culture gel. And that's probably the part that's probably least apparent to investors, but it's actually most important if we're going to go execute to the strategy. Maybe what are the top two or three priorities for Intel's leadership over the next 18 months or so? Which of these items would you say are truly make or break from an execution perspective? Yeah, it's a really good question. I think when Lip-Bu rejoined, he kind of highlighted four key priorities, which have kind of morphed into our strategy. One is to sort of fix the core x86 business. The other is to really solidify our strategy in AI. It's also to get Foundry off the ground, and then it's to improve the balance sheet. I would say actually this quarter we've done an excellent job on improving the balance sheet. I'm sure you'll have some questions on that. You know we started the quarter by monetizing about $1 billion worth of our Mobileye ownership, excuse me. We'll end the quarter with about $3.5 billion closing of the Altera deal to Silver Lake. You saw the announcement with the U.S. government, which added $5.7 billion. We did a $2 billion SoftBank equity investment as well. We've massively, I think, significantly improved the liquidity, and we've kind of hit on that one. On the Foundry one, I would actually say that's probably one of the more critical ones. We've talked about our desire to get 14A right. We talked a little bit about where we are in the development phase of 14A and trying to get external customers there. That's obviously, I think, a critical area for us to create value for our stakeholders. You look at the x86 business. I think we've got a strong product portfolio in PCs. We've got a couple of holes we've got to fill on the desktop front. Quite frankly, we feel confident in the roadmap. We'll have a refresh of Arrow Lake next year, which will help start the process on the desktop side. We'll conclude that with Nova Lake when we launch late next year into 2027. Lastly, on the AI front, as we talked about on the last earnings call, we owe it to our investors to give a deeper sort of view of where our AI strategy is going. I don't want to get too far in front of that. We'll probably have more to say on that when we report our Q3 earnings. That's a lot of stuff. Yes. But, which is the most, one or maybe two? I think this is part of the strategy here. We've got a lot that we want to accomplish. As you saw after the close today, we made some key leadership changes. I think Lip-Bu does have a lot on his plate, and he needs to find the key people in his team around him to help go execute. Just to kind of, at a high level, talk about some of the announcements we made, I'll start with Intel Foundry where really Naga has been running that organization since about July of last year. He was brought over from Micron to run all global manufacturing. Earlier this year, he was also put in charge of TD. His role has now expanded to also include the IFS interface with external customers. I think he's a key leader in Lip-Bu's executive team. In addition, we've made a couple of new hires as well. Kevork has come over from Arm, where he was a Senior Vice President of Engineering, to actually run our data center group within Intel products. He's got, I think, a very strong resume on SoCs. Before Arm, he was at Qualcomm. Before Qualcomm, he was at NXP. Lip-Bu had talked about the need to bring in some external talent to really help shore up our data center roadmap, and I think Kevork goes a long way of doing that. Internally, he's elevated Sachin to run our AI business. You remember in Intel products, we have two subsectors. We've got DC AI, data center and AI, Kevork and Sachin. Overall, on the client side, Jim Johnson, or JJ, as we call him internally, has been elevated to the permanent role of client. I think this is an important moment in time because I think Lip-Bu finally has his entire leadership team around him to go off and drive the strategy. As you point out, there is a lot that we need to get done, and it's hard to say what is more important than the other. I think critically, the 14A story is the one that probably drives the most incremental value for our outside owners, and we want to get that right. Fair enough. I want to get to that too. Yep. You talked about the government stake, the 10% stake. Is that a financial investment purely? More importantly, is there any kind of strategic impact to that investment in terms of shaping what you're likely to do or not do with respect to Foundry specifically? Yeah, the investment by the U.S. government, they don't have a board seat. They've agreed to vote their shares in line with management recommendation. From that perspective, it is financial. Having said that, even before the government owned a share of stock, they were a critically important stakeholder in the room around what they do on tariff policies, around what they do on exports, clearly the CHIPS Act as well. When you step back and look at the transaction we did with the U.S. government, we think it was a good transaction for us and for our shareholders. We think it was a good transaction for them and for the taxpayer. I think for us, going back to one of the key priorities of improving the balance sheet, you know this transaction with the government adds a lot of permanency to the capital around the CHIPS Act. And so, when the CHIPS Act grants were announced, we had announced that we won up to $7.8 billion of grants from the U.S. government. We had collected on about $2.2 billion of that by early January of this year. There was $5.7 billion that was still outstanding. There was some uncertainty as to whether or not we were going to be able to collect on that. Doing this agreement with the U.S. government took that uncertainty away. It put $5.7 billion of cash on the balance sheet, which we got about two weeks ago. That helped. There's also the $3.2 billion from the Secure Enclave, which is also part of this agreement. From that perspective, for our owners, relative to being able to fix or improve the balance sheet, we thought that this was a very good transaction. It also incentivizes, I think, the U.S. government to the same degree that our investors incentivize. The U.S. government now owns stock at $20.47. One of the critical KPIs that they're going to look at is the success of this deal. Is the stock price higher or is the stock price lower over time? We think that's a good partnership for them to have. Can you also discuss your sort of broader capital plans, including the SoftBank investment you announced? Help us understand what's the purpose and the use of these funds. Is this purely to shore up the balance sheet? Is it to bolster the dry powder for the Foundry business for prospective acquisitions you might make in AI or otherwise? Maybe just talk about the broad sort of plan from a capital perspective. Yeah, I think the equity that we raised in the quarter clearly helps to improve operational flexibility. That said, I do think that it's really about improving liquidity and being able to pay down maturities as they come due. We have about $3.5 billion of maturities coming due this year. We will use the funds to actually pay that down. In addition, I think we've got about $2.5-$2.8 billion for next year. Think about this cash as helping us pay down maturities as they come due. I think clearly, as you think about M&A, especially relative to our AI strategy, Lip-Bu has talked about both organic and inorganic paths to kind of drive that strategy inorganically through partnerships and potentially through tuck-in M&A. Clearly, having a stronger balance sheet helps to go off and execute that. How do you think about Intel's business portfolio in the long run? How important is software to the strategy? What capabilities does Intel need to add? How do you plan to get there? When you look at our business portfolio, I would break it down into two broad categories, Intel products and Intel Foundries. Both are relatively unique assets. On Intel Foundry, you know we're one of two, maybe three companies on the planet that can continue to prosecute Moore's Law. Quite frankly, we're the only one that does hardcore R&D in the U.S., which I think is an added benefit to all of our stakeholders. On the Intel product side, I still think that the x86 ecosystem is a very powerful ecosystem, both on the client and on the server side. I think as we execute to the strategy, both sides of the ledger will bring value to our stakeholders. Maybe shifting to your AI strategy for a second. You mentioned Rack Scale Architecture products coming from Intel at some point soon. Do you have all the necessary capabilities for that to happen? Are there other capabilities you'd like to have? How much would you expect you to use M&A to get there? Yeah, it's a good question, Jim. As I said earlier, on the last earnings call, we said that we would come out with a more fulsome sort of understanding of our strategy over time. I don't want to get too far out in front of that. What I would tell you is there's probably two prongs to the strategy. There's a tactical prong of repurposing assets that we have internal today to try to get a relatively modest, quick footprint into the market. That, to me, wouldn't be all that interesting to investors. I think there's a longer-term strategy that Sachin and Lip-Bu are driving about what assets that we need to really go off and be disruptive in this market. You know we've talked about having a focus more on inference versus training, having a focus on agentic AI and physical AI. I would ask you to be a little bit patient, especially as we head into earnings. We'll have more to say. I did have a question about sort of merchant versus custom solutions and that. Maybe I'll table that one. Maybe just ask you, broadly speaking, big picture, five years out, what role does Intel want to have or do you want to play in terms of the AI accelerator market? I think we want to have a meaningful presence over the next five years. I think in general, the way I would think about it is fixing our core x86 business probably gives us the opportunity to have a consistent growth business in that 3% - 5% range. I think Lip-Bu's aspirations are higher than that. I think to achieve those aspirations, we really have to have a bigger footprint in AI. We think that our x86 ecosystem brings value to that market. We think we can find ways to be disruptive, especially in inference around power and agentic around power as well. So, stay tuned. A couple of kind of short to medium-term business trends. Looking at the Q2 results, I think your revenue performance was really positive relative to what most investors thought. I think clearly margins of possibility are still a little bit challenged. Maybe talk to us about your thoughts around the things that need to happen for your gross margins to improve from here in your visibility and confidence level that those things are actually going to happen. Yeah, it's a good question. I'd point out that absolute margins, you're right, they're challenged. We did do better than we guided to in the second quarter, especially if you exclude some of the one-time charges. We do think we're on the right trajectory, albeit we still have a lot of work to do. Quite frankly, there are two drivers to getting better margins from here. One is to improve the competitiveness of our products. The second is to get 18A ramped successfully. One of the things that we've talked about is the move from Intel 7 - Intel 18A is very accretive to profitability in Intel Foundry. It is accretive to overall Intel corporate profitability. We're excited that we will be launching our first SKU of Panther Lake by end of year. We'll be ramping multiple new SKUs of Panther Lake coming in the first half of next year, and that will help on the profitability side. I think one of our biggest challenges this year is a lot of that revenue growth is coming from products that aren't as cost-efficient as they could be, specifically on the AI PC, with things like Arrow Lake and Lunar Lake, where we're outsourcing most of those tiles and paying that external foundry margin. In addition, with Lunar Lake, we've talked about the embedded memory, and that also is causing some gross margin headwinds. Those probably peak in the second half of this year. As we go into next year, and as we talked about on the earnings call, we think you should think about incremental gross margins in that 40% - 60% range as we continue to execute. Yeah. In terms of OpEx, what's the philosophy sort of near or medium term? You know, given that you've kind of addressed some of the capital structure concerns, is cutting OpEx from here a desired outcome or not? What are the areas that you're prioritizing? Yeah, we've done a lot of heavy lifting on getting OpEx into the right area. Remember, we guided OpEx this year to $17 billion. For next year, we guided it to $16 billion. I think those are both kind of the right numbers to think about. I think that the gross reductions we're seeing in spending are more significant than the net. There are areas that we need to invest, whether that's the AI roadmap, reinvigorating data center, especially around multi-threading, or investing in the development of Intel 14A. Those are all pretty critical areas. Now, having said that, I'll remind you that longer term, it's probably a good, healthy model for a semiconductor company to be running about 20% - 25% of their revenue as the percent of OpEx. Against sell-side street numbers next year, $16 billion doesn't get us there. It's the right number for us to target for next year relative to the investments we need to make. We need to do better over time, either through revenue growth or more efficiency gains. Fair enough. Maybe kind of shifting to Intel Foundry Services for a second, can we maybe kind of address CapEx upfront? What's the base case you're planning to in terms of the 2026 CapEx? Can you maybe share any kind of guideline in terms of your thinking about steady state capital intensity once 18A is fully ramped? Yeah, again, another good question. I think we guided CapEx this year to a gross basis of $18 billion. We've talked about CapEx for next year being down. On the earnings call, we also talked about maintenance CapEx at around $9 billion. I think that got misinterpreted that CapEx could be down a lot next year. I don't think that's the right way to think about it. I think we will be down, but I think modest is probably the right neighborhood of down to think about because we still have a significant amount of the 18A capacity build in front of us. Yeah. One thing I'll remind you is this year's 18A capacity is really being put in place to support Panther Lake. Panther Lake is only a notebook part. As we think about next year's CapEx, it's really going to be to support Nova Lake, Clearwater Forest, and Diamond Rapids. Just Nova Lake itself, being both a notebook and a desktop part, has pretty meaningful implications for the amount of wafer starts that we need on 18A. I think a lot of that is still ahead of us. Now, I think longer term, we haven't put out a capital intensity. One of the clear things that Lip-Bu wants to signal is he is going to be extremely financially disciplined when adding capacity. Some of the disclosures we had in our 10Q off of the Q2 earnings around 14A were really there to punctuate the fact that we're just not going to add capacity until we see demand and the need for it. I think that's a shift from how we've been operating in the prior 3+ years, where we did have a little bit of a strategy of if we build it, they might come. Quite frankly, I think the most tangible sort of example of the overspending is over the last three or four years, our assets under construction went from $20 billion - $50 billion. That's, to me, a clear indication that we bought ahead of demand. I think one of the things that Naga is really focused on is being more efficient and taking that number down. For your 14A node, your message, I think, has been you need to see volume commitments from external customers before you put new CapEx dollars to work. Maybe speak to what you believe customers are focused on for 14A in terms of technology proof points and volume readiness. Yeah, just so there's no confusion, we are all in on `Intel 1 4A development. That said, we did talk about on the earnings call that in order to generate a good return on our owners' capital, we need not only demand from Intel products internally, we also need to sign up a significant external customer to make the node work. If you look at what those external customers are looking at, it's really two things. It's PDK readiness and maturity, and the yield curve. We feel very good about where we are in Intel 14A development. The way I would think about it is, simplistically thinking, there are three phases to any new node. There's the definitional phase, there's the development phase, and then there's the high-volume manufacturing phase. I think Intel 14A is a very different node for us. At 18A, in the definitional phase, we were really only working with Intel products. All of the choices we made on the 18A node were really to optimize for the internal customer. We didn't really engage external customers until we got to that development phase. To be clear, for some customers externally, the fact that it was optimized for Intel products didn't make a difference. For others, it did. The big difference on Intel 14A is right in the definitional phase, we are actively engaged with external customers to define the node. We're also working with Intel products. Quite frankly, what that really means is Intel 14A from the get-go is more suitable for external foundry customers. It also means we're getting earlier, more, and better feedback from those external customers. We've talked about some of the hard design choices for Intel 14A that our customers will need to make, probably happening in the second half of 2026, going into the first half of 2027 as far as proof points. Given the interaction we're having today, we'll have a good sense of the trajectory of our success probably well before that. From an investment community perspective, what are the key developments and milestones that investors should be tracking to determine whether Intel 14A is on track or not? The role of the Foundry is to make their customers successful, not to market their customers. I'll leave it up to you guys to do your channel checks and find out how we're doing. In general, it really is about PDK progression, the yields, and progression of test chips. I'll remind you, in April of this year, when we had our second annual Intel Foundry Day called Direct Connect, we did mention that early customers on Intel 14A have an early version of the PDK. I'm sure as we report earnings, we'll give you sort of 90-day updates on how that is going. I'll let you guys run your channel checks elsewhere. OK, very good. I'll get into that right away. Maybe can you detail your strategy on your strategy to make Intel Foundry profitable? What are the milestones you need to think about internally? Is there any shot at 2027 break-even being a base case? Yeah, Naga is still driving the organization to be a profit break-even exiting 2027 on a run-rate basis. The biggest driver of that is ramping the Intel 18A node. We've been pretty transparent that the Intel 7 node just was not a cost-competitive node. As you look at it through the lens of Intel Foundry, the move from Intel 7 - Intel 18A, ASPs per wafer for them will go up three times faster than their cost. Just driving more volume through the fab on 18A is a pretty profitable sort of dynamic for Intel Foundry. It's mainly on the back of Intel products. We don't need to see a lot of external foundry revenue to get the profit break-even exiting 2027. That's still the goal that Naga and his team are driving towards. Yeah. OK. I think on your last conference call, I think you sort of opened the door to your product groups having a little more latitude to use foundry partners for an open-ended period of time, which I think is a bit of a tone change. Maybe help us understand the pros and cons of keeping external foundry as a separate business and even the idea of keeping manufacturing in-house. What are some of the options the board may be weighing? Is there a scenario where Intel exists as an independent design company only? I think we've been on a journey over the last several years to create some separation between Intel products and Intel Foundry Services. It started with just financial transparency of actually giving Intel Foundry Services their own P&L for the first time ever. I think that was also beneficial because it helped that group understand what they can go after to actually improve profitability. Quite frankly, it did the same thing for Intel products. Like in the old model, when most of your operating profit is dependent upon allocated cost, it's really hard for you to go figure out what you can actually directly affect to improve your own profitability. We thought there was a benefit to both the Intel Foundry Services group and to the Intel products group. Obviously, the benefit of an IDM model is margin stacking. We've talked about that. You don't have to pay that external foundry margin. You actually get to capture it. Having said that, I think it's important to note that Lip-Bu is a best products win guy. It doesn't matter where those wafers need to come from to generate the best products. To your earlier point, Intel products absolutely have more autonomy to use external suppliers if they think that is the right choice. I'll remind you, given our product roadmap today, we know over the next couple of years that we'll be bringing more wafers in than we have this year, only because Arrow Lake and Lunar Lake, we know, are fully outsourced products. Panther Lake starts the process of bringing that back in, and that will help improve the profitability of Intel Foundry Services and overall Intel corporate. I think longer term, maintaining the optionality of creating more separation between the two businesses makes sense. There are obviously external foundry customers that we're engaged with today that would like to see that over time. We have to be a little bit pragmatic about it. The first step is to improve the profitability of Intel Foundry Services so that it can stand on its own and not be wholly dependent upon the cash flows in Intel products. This will be a multi-year journey, but I think we have preserved the optionality to go down that path if we think it's the best way to create value. Yeah. If we think about the data center and enterprise business for a second, how are you seeing the server CPU market in terms of growth rates and key drivers of that? Are we thinking about replacement cycles, core count, pricing, things like this? Is that the main driver of it? How do you think the growth is structurally going to track for the next few years? All of that is definitely in play. I will say that through Q2, it is obvious that the TAM and, quite frankly, our relative market share both held up better than we thought in the core server business through the calendar second quarter. A lot of that, I think, were companies that delayed refresh in prior years, given the AI acceleration boom. I think also there's probably some argument to be made that AI is just driving higher compute demand across the spectrum. I think also that we're well positioned in some of those AI servers for the head node. Now, as we think about this business longer term, we really think about it around core growth and ASP per core. Our long-term models are projecting a unit market for servers that's plus or minus flat-ish. We think that could end up being a conservative view. We think that's the prudent view as we try to manage the business. It really becomes, how fast do you think your core count is going to grow? What do you think ASP per core is going to do? I think historically, we've seen core counts grow in that 20% - 25% range. I think the law of large numbers means that probably slows somewhat, I think, mid-teens. The real, I think, driver is going to be what happens to ASP per core. We still think it goes down, but at probably a slower rate than it historically has. That should drive some pretty healthy revenue growth for the market over time. Our number one priority is to improve our product competitiveness. As I said earlier, I think our share has held up better than we anticipated year to date. We're still not happy with where we are in the marketplace. I think we've been pretty transparent that we have some holes to fill around multi-threading. I think on the single-threaded performance, we feel really good about our position. We've got some work to do over the next couple of years. If you think about your server CPU portfolio and roadmap, help us understand what customers are asking of you. What are the customers' conversations like? What are they asking you for in terms of future product developments, both on the hyperscaler and enterprise side? Better TCO. I apologize. It's not a glib answer or a short answer. That's exactly what they're looking for. It comes down to performance per watt. It comes down to their cost of ownership. I think in general, one of the things that Lip-Bu is really trying to drive inside of the organization is much more of a customer-focused, engineering-centric culture. I think one of the things that we could afford to do when we had a 2.5 year process lead over everyone else and dominant share in the markets we participated in was not listen to our customers closely enough. I think that that has come back to hurt us a bit. I think Lip-Bu is changing that real time. Yeah. On PCs, I think the market has been a little bit better for a little bit longer than many people thought. How would you describe sort of how the company views the demand outlook for the next 12 - 18 months, both from core PCs and the contribution of AI PCs, both in the enterprise and on the consumer side? Yeah, so again, I'll repeat what I said about the PC market relative to the server market. Year- to- date, demand has been stronger than we had anticipated. We've been concerned in the first half of the year around tariff pull-ins. I think as we look at a sell-in basis, it's hard to find evidence that there has been significant tariff pull-in, i.e., inventory levels on a sell-in basis actually look relatively healthy. Healthy for us is 6-8 weeks. Where we have a little bit less visibility is just sell-through. It's unclear to us whether or not end users decided to buy more PCs in the first half of the year in anticipation of costs going up because of tariffs in the second half. It really, I think, is the foundation of why we guided the second half subseasonal in our business, both for Q3 and for the second half. Having said that, there are things to be positive about in the PC market. We're still very early on the AI PC trend. We'll be less than probably 30% penetrated on AI PC this year. We don't think the crossover really happened until the second half of next year. That could be a very powerful trend for multiple years to come. We're also not as far along on the Win 11 upgrade cycle in corporate as we were at this time on Win 10. While we're being prudently conservative about the second half of the year, there are still strong underlying growth drivers over the next 12 - 18 months. How do you think your market share position in PCs kind of tracks out over the next year or so? It seems like it's been also a little bit better than people had thought. I'm curious how you think that trends. Yeah, in general, we're still shipping relatively 7 out of every 10 PCs. Our x86 competitor ships 2, and the Arm competitor ships about 1. We feel pretty good about our position. Clearly, I think this year we've seen some challenges at the high end of the desktop market. As I alluded to earlier, we're going to see a mid-cycle refresh of Arrow Lake coming up, which should help stem that a bit. I think as Nova Lake comes out at the end of next year into 2027, I think we're going to have a leadership position across the board on desktop. On the notebook market, we feel good. I think Lunar Lake absolutely was a great product for this year. I think it proved, or I think it bunked the myth that x86 can't have 20 hours' worth of battery life. I have my own Lunar Lake. I need to charge it about once a week, and that's fantastic. From that perspective, I think when Panther Lake comes out, we just build on that momentum. I'd also be remiss not to point out that our enterprise sort of ecosystem lock-in is still pretty strong with vPro and the like. If there is an acceleration of the Win 11 upgrade cycle, we should be well positioned to benefit from that. Great. I think with that, we're almost out of time. I want to thank John for being here. I appreciate you all being here as well. Thank you. Thanks, John. Thanks, Jim.