Intel’s CEO Says Using Competitor’s Process Tech In Intel Fabs Is an Option
Intel’s CEO Bob Swan isn’t a stranger to conducting interviews, but his appearances have traditionally been limited to the financial community. To our knowledge, he hasn’t held a question and answer session with the technical press in the past. That changed last week when we participated in a roundtable discussion with Bob Swan covering a wide range of topics. You can read the full transcript of the conversation below (after our commentary).
Perhaps the most interesting revelation from the discussion comes in response to one of my questions – is it possible that Intel could license and use a competitor’s process node technology, but inside its own fabs?
Bob Swan’s response, that it is a possibility, would have seemed inconceivable at the beginning of last year. However, Intel’s recent revelation that it will outsource some leading-edge chip production in the wake of its delayed 7nm node, and the fact that the company still hasn’t decided just how much, what, or where it will outsource, has led to serious questions that have resulted in bruising stock valuation declines for the world’s top chipmaker.
These same questions recently led an activist investment hedge fund to question the company’s plan forward, asking Intel to consider spinning off its fab operations.
For many, Intel outsourcing its chip fabrication is akin to Ford contracting with Chevy to make its cars, but the prospect of Intel spinning off its fabs entirely seems unlikely. While Intel currently outsources ~20% of its production, the company has historically limited its outsourcing to low-margin and low-complexity products on trailing-edge process nodes. However, in the wake of its 7nm delay, Intel has signaled an increased willingness to embrace third-party foundries to access their leading-edge nodes, thus expanding its existing use of third-party silicon to its core logic components (a first).
But there are complexities wrapped within complexities. Intel’s native chip production capacity has helped it to largely avoid the shortages we’ve seen with other chipmakers, like AMD and Nvidia, in the wake of the global pandemic. Intel’s tightly-controlled supply chains are also a strength that has helped it combat the waves of shortages that have impacted all facets of semiconductor production, such as secondary componentry like substrates and power ICs.
Outsourcing doesn’t seem to be the best long-term solution to Intel’s woes, and it may not be feasible on a large scale: Intel has more than twice the semiconductor output of TSMC, which is largely thought to be Intel’s presumptive partner for third-party chips. Given that TSMC is already capacity constrained, it obviously wouldn’t have enough output to satisfy Intel’s incredible volume without making massive long-term investments of its own. Besides, Swan recently said that even though Intel will now engage third-party foundries as strategic partners, it will continue to develop its own leading-edge nodes and has deployed a “fix” for its own 7nm node (though that fix has led to an untenable delay).
Unfortunately, as Swan expanded on in response to my question, outsourcing some chip production could expose Intel to possible supply issues, and naturally, outsourcing production could lead to lower margins.
But there’s possibly a middle ground: There certainly is precedent for a fab to license another company’s process node technology for use in its own fabs, and according to Swan, that is an option for Intel. GlobalFoundries employed a similar strategy when it licensed Samsung’s 14nm process technology back in 2014, and it’s possible the company would be open to a similar deal in the future. Conversely, Intel could also approach TSMC with a similar request, so anything remains possible.
This middle ground would allow Intel to maintain tighter control of as much of its supply chain as possible, preserve margins as best it could, reduce investment in outside firms, and perhaps divert more of its own internal engineering resources to its future nodes. It could also help the company maintain its advantage of architectures that are tightly coupled with its process tech. Perhaps by producing the chips in its own fabs, Intel could use the yield learning curve to its advantage to increase profitability.
There are precious few semiconductor toolmakers in the industry, like ASML, which grants at least some level of standardization to the leading-edge chip manufacturing equipment. That could help ease the implementation process if Intel chooses this path.
Swan has previously stated that the company would provide some clarity around these lingering questions early this year, which most project will happen at the company’s next earnings call on January 21. His statements in the question and answer session below give us a deeper understanding of some of the considerations behind the company’s looming decision. If one thing is clear, it’s that Swan appears committed to Intel remaining an IDM, even though the company’s definition of being an IDM is changing as it looks to sidestep the issues with its own process technology.
The format for the routable was simple; Bob Swan opened up with an overview, after which the invited participants were allowed to ask questions. Time constraints limited us to one question apiece. The transcript follows:
Bob Swan, Intel CEO: First, I will start with the market dynamics in terms of how we see it. I would say it’s a good time to be in semiconductors. The digitization of everything seems to be accelerating, and computing is everywhere: It’s no longer just our PC or server. Everything seems to need high-performance compute, and the PC is essential once again. So, the dynamics of the market are extremely favorable. To capitalize on these data-centric transitions requires massive transformation. Not just for all of your clients and obviously for us, but a massive transformation so that we’re adjusting, adapting, and developing the new technologies that are going to enable these transformations like 5G, AI, autonomous, and the autonomous edge.
In some ways, fortunately, and in some ways, unfortunately, we’re not the only ones that have noticed this massive opportunity in front of us. As you know, competition is intense – they never sleep. And that means we have to be on top of our game, which we have every intention to be as we close out last year and enter this year. So, we’re excited about the competitive landscape. Sometimes I wish it would be less intense, but I think it makes us stronger along the way.
The market overview is relatively attractive. Strategically, what we’ve been talking about for the last couple of years, in terms of our role in this market, is to be the trusted performance leader that takes this insatiable appetite for data and makes that data relevant and actionable by analyzing it, storing it, moving it faster and faster, so that whether it’s a business or a consumer going through their own digital transformation our technologies are there to make the data that everybody is getting after increasingly relevant.
We know that for us to achieve our dreams, it requires us to transform ourselves on three fundamental dimensions. One, from, as we characterize it, the CPU to GPU to XPU. As the industry evolves, as workloads evolve, having a variety of different architectures enhancing our core CPU, and also adding additional architectures, is increasingly important. Secondly, from silicon to platforms. Not just the hardware, but how do we couple the hardware and software with other technologies to build platforms that can delight our customers.
The third area for us has been a transformation from what we’ve characterized as the traditional IDM [Integrated Device Manufacturer] to a more modern IDM. This is not to get rid of the IDM model; we think it’s a unique advantage of this company, but we know the industry has evolved quite a bit. We know that by leveraging design, disaggregation, and packaging technologies, the IDM of the future is going to be a little bit different than the IDM of the past. And that’s the third leg of our transformation.
And then, to bring it all together, our intensity and focus on execution is tightly coupled with reenergizing the culture of the company. So we’re moving in a much more nimble way as we see opportunities to grow and innovate, as opposed to protecting what we’ve built in the past.
A year ago, I flagged three fundamental areas that I thought were critical for us to improve execution. One was the capacity to ensure that we have the capacity to meet the demands of our customer base. The second was, once and for all, the ramp of 10nm. The third was to increase the rate of innovation on this multiplicity of architectures.
As we enter 2021, we’ve built some really good momentum. Obviously, we have a lot more to do, but we’ve added over $20 billion in revenue over the last five years, and we’ve been chasing it. And chasing it is disruptive. We exit the year having essentially doubled our capacity over the course of the last couple of years, and we’re going into 2021 with a lot more capacity in place, both for 14nm and also for 10nm.
The second area is 10nm itself, which we finally launched last year. During the course of the year, going to the second generation, or SuperFin, was the biggest internode enhancement in transistor density ever for us. Coupling that with our packaging technologies, bringing them to life, was a really important aspect of our execution on 10nm.
And then third is products, just the rate of innovation. And during the course of the year, whether it was at the beginning of the year with Lakefield, 5G SOCs, FPGAs, discrete graphics, oneAPI, or the amount of products we launched during the course of the year, we had really good momentum. And I would say that maybe most importantly, the Tiger Lake launch during the course of the year ramped on the new node faster than we anticipated. It grew better, and the adoption and designs by our customers were more than we expected. So, we enter the year with that momentum behind us and the opportunity to really scale 10nm as we come into the year.
The last thing I’d say about the execution as we exit the year – we launched the Ice Lake server product, qualified it at the end of the year, and started production of the Third-Gen Xeon Scalable processors. We are ramping [production] in the first quarter. As you know, that was an important launch to complete the portfolio of products that we have, from Lake Field at the beginning of the year, all the way to server at the end of the year. It’s good to have that launched and ramping production as we begin this year.
So that’s a little bit about where we’re headed strategically and the momentum we have entering the year. And then I’ll touch just quickly on the things you’re going to hear about at the press conference on the 11th. I mentioned the Ice Lake Xeon Scalable. Next year the client business is going to show more advances in industry firsts for virtually every type of experience in every segment of the PC market. During the year, we’ll have four families of processors from entry to premium, a really good lineup of products during the course of the year. That lineup will include some real desktop innovation. Gregory Bryant or Chris will showcase two new technologies coming to market in 2021, including the 11th-Gen Core desktop processor, Rocket Lake, and our next-generation processors, codenamed Alder Lake, which from our vantage point represents a significant breakthrough in the x86 architecture. It’s our most power-scalable SoC for both desktop and mobile processors. We’re really excited about the innovation on the desktop, and between server, client, mobile, and desktop, we have a really exciting lineup of things as we enter the year.
The second thing we talked about is silicon to platforms, which will also enhance our platform offering this year. We’ll launch our first-ever 11th-Gen Core Chromebooks based on the EVO platform. We’re also going to debut the Intel EVO vPro platform as well, which is going to feature the highest-performance business PC platform with the most comprehensive hardware-based security.
So, we’re very excited about going beyond the silicon to the platforms, and not just for Chromebooks. You’re going to hear from Amnon about how Mobileye is expanding its offering as well. They’ll talk about expanding autonomous vehicle efforts globally, with four new testing locations in Detroit, Tokyo, Shanghai, and Paris in the first part of 2021, and basically demonstrating the ability to have that kind of rapid global deployment by leveraging the proprietary crowdsourced mapping technology called REM (Road Experience Management).
So, just to wrap so we can have a dialogue, I can stop talking. The demand for compute is pervasive. It’s not just that the PC or the server, but at the network, at the edge, in the car, in the retail store, and the industrial environment. And that incredible demand for compute on billions and billions of devices is a rich opportunity for us to expand the role that we play in our customer’s success. And we feel like we had good momentum in 2020 and entering the new year, and we’re excited to demonstrate a bunch of new innovations. So, let me pause. Sorry for the long-winded rant about the exciting stuff we have. But I just want to get that out there so we can have a discussion.
Patrick Moorhead, Moore Insights & Strategy: Can you give us the latest and greatest on 10nm?
Bob Swan: Yeah, well, maybe three components. First, the [production] capacity. I think, as you know, we’ve got three high-volume fabs ramping now, so capacity for this year and the years [following]. They’ve gone out a couple of years – we have the capacity in place, first and foremost.
Second, in terms of yields, unit cost is obviously driven by wafer costs. Front end and back end yields have been progressing very well, almost quarter-on-quarter during the course of 2020, and they ramped even faster in the fourth quarter of the year. So, unit cost driven by those three factors continues to come down, i.e., all else equal, our gross margins get better and better as we ramp it.
And then the third is obviously, I repeat myself, but it’s not just the process technology, but what products are you putting on it? I think we ramped everything during the course of last year, and the Ice Lake product was the last one to ramp. In 2020 we said, it might have been architecture day or maybe in our third quarter call, that volume on 10nm this year was 30% higher than we anticipated beginning the year. And obviously, Tiger lake was a big driver for that. But we felt great about the design wins and our ability to ramp across all three facets as we exited [the year].
Paul Alcorn, Tom’s Hardware: Right now, the industry is going through a lot of shortages. One of the key benefits of Intel is that it manages its own manufacturing supply chain as an IDM. But you also use a lot of external suppliers for some components, and due to COVID and political reasons, expanding the scope of third-party supply could present some challenges. How do you plan to move forward in a future where you blend more of your product lines with a mix of third-party derivatives, and the supply and economics challenges that brings?
As a close question to that, is there a scope for Intel to enable a licensed derivative of another foundry’s process node technology? Put in another way, use someone else’s process technology in your own foundries?
Bob Swan: Great question. I think first on your first point; it’s just access to supply. You may remember it was a year ago that, collectively, we all looked into 2020 thinking, ‘well, PC might be plus one or minus one [percent].’ It kind of feels like it was about a century ago, but the demand profile wasn’t that great, and all of a sudden, COVID hit. The immediate reaction was, ‘Oh my god, the second half is going to fall off like a rock.’ I think that was the industry’s response to it, and lo and behold, the second half grew like crazy. Not only grew, but the mix of the products changed dramatically as well. The whole ecosystem was scrambling for how to keep pace not just with increased demand but the changing mix of desktop and mobile devices. So, we all had to struggle with that.
I think, for us, you flagged it; this is one of the massive advantages of an IDM. We’re still dependent on other commodities but being able to control your own destiny and be the allocator, as opposed to the allocatee, is a big deal. So, when we think about the challenges of just supply constraints in general, but even as we think about design disaggregation and where we may leverage other third-party foundry’s capacity, in any discussions along those lines, you can imagine that we are asking, ‘even if we don’t make it, how do we preserve the advantages of IDM?’
More simply put, how do we leverage our size and scale with the entire supply chain to get preferential treatment and be first in line when allocation decisions are being made. We’ve got a lot of practice doing that with commodities, and now as we think and evaluate whether to expand in a disagg [disaggregated chip design] world with more utilization of third-party foundry, maintaining and securing the advantages of an IDM is important to us in considerations.
I think the second part of your question was, is there a time in the future that we would use somebody else’s process technology inside our manufacturing environment? And I would just say that…possibly. And, I think strategically for us, we know that the ecosystem has evolved quite a bit over the last ten years. If there are ways or opportunities for us to leverage some of the advancements of the industry in new and different ways, I think it’s going to be very front and center for us to capitalize on industry innovations; we don’t have to do all the innovations ourselves.
And that means we may outsource more; it means we may use more available third-party IP, it means we may make stuff for others, not just, i.e., be a foundry ourselves. And is there a scenario where we could be using somebody else’s process technology in our fabs? That’s possible. The key is, as the industry evolves, how do we leverage the innovation? Not just within our four walls, but the innovation happening in the industry as a whole, and be very flexible and adaptable to take advantage of those [innovations] along the way. I hope that helps.
Mark Hachman, PCWorld: I think you alluded to the competition earlier, and I wanted to focus on that for a second in the Core space and the PC space. Right now, if I’m a PC vendor, I obviously have the choice of buying from Intel, I can buy from AMD, or at least hypothetically, I can at least roll my own ARM processor, or at least consider buying one from a third party. Right now, what is the competitive advantage of the Intel Core processor? Why should a PC partner buy from you as opposed to the competition?
Bob Swan: Yeah, I think it’s a great question. Again, this goes back to my intro comments; in the good old days, there wasn’t that much. But now we have the good old days where competition makes us stronger, makes us move faster. First, it starts with the rate of innovation, and then broad-based product offerings. Having products from intro to premium, having products for business, for consumer, for notebook, for desktop, and then that annual predictable cadence of leadership products. That’s what we’ve been known for over time.
Secondly, and back to the IDM, the knowledge in the customer base that we are not going to constrain your growth. I.e., I’m going to have too much capacity before I’ll have not enough capacity, and I don’t have to wait for somebody else to allocate to me. I’m going to be the allocator, and you can count on us to have the inventory on hand to meet your spikes in demand.
And then the third is knowledge. Our ability to work with other engineering teams to almost co-optimize the products that they are working on. It’s the technology dialogues that we have with the OEMs, in particular, so that we’re part of their design team, not just an order taker for their product needs.
Dave Altavilla, HotHardware, HighTech Vision and Analysis: I’m just wondering if you could provide an update on the progress of higher-end Xe GPU variants, how that’s going and how you see perhaps it will compete in the market for various GPUs, whether it’s data center or gaming?
Bob Swan: We’ve had a couple of goes at this, as you know. When we started over a couple of years ago, rather than starting from starting from scratch like we did with Larrabee, the idea was, how do we start with integrated graphics and enhance integrated graphics, but do it in a way that migrating from integrated to discrete is much easier to leverage as the same design.
You saw the progress we made on integrated graphics last year. And we’re very excited about the launch of DG1. I actually don’t know if we’ve mentioned or given a date for when DG2 two is [the launch] and when we launch the data center. But clearly, our intentions are, in this CPU to XPU world where there are multiple architectures required, whether it’s gaming or whether it’s in a data center, we’re going to invest to extend our graphics capabilities from entry-level all the way up the stack over time. We expect to have real innovation rolling out during the course of the next couple of years.
Ian Cutress, AnandTech: Given Intel’s position in the market, and the outlook for 2021, where you currently stand, what matters more to Intel, the financial side or the technical side?
Bob Swan: Technical. As you know, they’re not decoupled: Rate of innovation, rate of products, capacity in place, investing in technology development – it’s one drives the other. We know that whether it’s capital, whether it’s continuing to increase R&D and compress the cycles, and when we get products out, that’s what drives our financials. It’s very tightly correlated.
In the scheme of things, it’s relatively easy to maybe save money. But we can’t save enough money to offset what incremental demand and volume do force across the spectrum of both architectures and where compute happens. So, the simple answer is technology. This is all about innovation relative to the other guy so we can get paid for [unclear], so we don’t have to share all the profits with the others in the ecosystem, which gives us the capacity and the fuel to, in turn, drive more technology and drive more innovation.
Ian Cutress: Intel has had a recent exodus of prominent key technical personal in 2020, such as Renduchintala and Keller. Where does Intel see the future of its technical expertise coming from?
Bob Swan: We have 70,000 engineers with technical backgrounds in the company. Our model for the longest time was to hire engineers right out of college campuses: It’s the input, and then we grow them within our system. But we also realize, particularly as we think of more architectures and how much the ecosystem has evolved for the last few years, we’ve been trying to both grow from within and bring outsiders in to try to get one-plus-one equals five.
As you know, Jim left for personal reasons. Murthy left, he’s a wonderful executive, but in so many ways, it was, ‘how do we get five engineers in the room when we’re making big decisions, not just one?’
And I think it’s the realization that we can get more diverse points of view, more technologists in the room when we’re making big-time decisions. Now we have a chance to see the Ann Kelleher’s on the TD [Technology Development] front, Keyvan Esfarjani, who ran manufacturing and ops for memory. Taking Ann’s spot, we have Raja Koduri on architecture and software. We have Josh Walden in the Jim Keller role now as we look to fill that role. We also have Dr. Randhir Thakur from Applied Materials on the global supply chain.
Having five senior, what I’d characterize as brilliant, engineers in that room gives us more diversity of thought on the rate of innovation and ideas. Then there’s a whole group of core hardware and software engineers that work for them, and I’m convinced we have the best talent, by a long shot, in the world. Our expectations are to continue to have the best talent, and we’ll get it the same way as we have recently – which is from the campuses and the competition.
Intel PR Representative: Bob, do you want to close it out?
Bob Swan: I would say, first, thank you. I appreciate you taking the time; sorry we’re so time-starved. I look forward to being back at CES in person, or seeing you at Architecture Day or Analyst Day. It’s been a tough couple of years. We’re coming out of 2020, and our guidance said it’ll be the best year in the company’s history. It’s the fifth best year in a row, and to your question, Ian, it’s all about innovation and execution. And we’re getting our mojo back as we exit the year. We’re excited about what we’re going to be able to do for our customers in 21. And I look forward to continuing the dialogue with you all. So Happy New Year. Thanks. Thanks so much for taking the time.
#Intels #CEO #Competitors #Process #Tech #Intel #Fabs #Option