Below is the transcript of a CNBC Exclusive interview with Intel CEO Bob Swan and CNBC's Arjun Kharpal. The interview was first broadcast on CNBC's Squawk Box Asia on 1 August 2019. The interview took place at the Rakuten Optimism conference in Yokohama, Japan. If you choose to use anything, please attribute to CNBC and Arjun Kharpal.
Arjun Kharpal: Bob I want to talk about this trade story with the US and China as this trade dispute continues to roll on. Semiconductors really been dragged into this dispute. What kind of impact has it had on Intel?
Bob Swan: Well first I would say that for the semiconductor industry as a whole the US has been a net net exporter for semi. So it's been a big source of innovation domestically in the US and we the industry at Intel in particular have been deploying that technology in the markets for a very long period of time. For our company in China, we've been in China for over 30 years. So we've built very strong relationships with domestic customers. We have very strong relationships with global OEMs that assemble in China and then ship in China or rest of world. And then we have a manufacturing environment and a lot of very talented employees. So the market has been very important to us for a long period of time and we've been very successful. So what we've been trying to do with the with the respective administrations is just encourage them to sit down and talk about, how do you create a more level playing field where trade is not impeded by tariffs, because we believe that tariffs ultimately just mean a tax on consumers or customers at the end. So we're pretty encouraged that they're back talking, but at the same time, in the meantime, you know the China market for us has slowed and it's had an impact on how we see the full 2019 playing out as demand signals overall economy in China slowed and then demand signals for semi equipment and for our business in particular slowed as a result of slowing demand.
Arjun: You mentioned of course the two sides are back at the table. So what would you ideally like to see out of a trade deal? What are the things that you would like to see included?
Bob: Well I think you know it starts with fair trade which is a function of ensuring that everybody has equal access to respective markets, number one. Number two that it's you can protect your IP. And number three that there's not imposed tariffs or penalties going either way that impede the flow of free trade. So those are the three things that we think are extremely important. And again we're encouraged that they're sitting down talking, but we hope that there could be some resolution soon.
Arjun: And in the interim what have you been doing to your supply chain. Have you been making any changes or any shifts in that?
Bob: Yeah I mean particularly with the global OEMs, because this as you know started over a year ago. And when we start to see tariffs for products coming out of China, the global OEMs and Intel working with them, is how do we adjust the supply chain dynamics to reduce the likelihood that the flow of goods are impacted by tariffs. So yeah it's been a very important relationship with the OEMs. How do we work through this equation, to have the de minimis impact on tariffs and impact our customers, and I think the industry and Intel is quite a very important role to try to mitigate the ultimate tariffs.
Arjun: And in terms of your customers in China, what response have they had to the trade dispute. Have they been stockpiling Intel chips or have they been taking any other actions to perhaps mitigate some of the impact?
Bob: Well for sure what we experienced last year was when the original first three tiers of tariffs came out last year, we noticed quite a spike in demand in the third quarter of 2018, from particularly the large cloud service providers in China, significant step function and demand. So clearly you know we have a wonderful leading position, our products and technology are embraced by the cloud service providers around the globe, and when they are in fear of not getting access, there was a real spike in demand last year. And the implications of that is now they're digesting, both domestic cloud service providers in the US as well as China cloud service providers. There was a spike in purchasing last year which is little normal for cloud service providers, but it was exacerbated in China, and now we're going through that and digestion phase of those significant buys last year.
Arjun: And is that really why we've seen that slowdown in the datacenter business there and do you expect a recovery in that business. What you're not seeing is a saturated market in cloud at the moment.
Bob: No no. I mean I think overall the ultimate end market trends that drive demand for our product are very strong. And in essence it's you know, this demand for data, and that's not slowing down. The data that's created is expected to grow by 25 percent a year. It's going to grow in the data center and at the networks, and the desire to get access to that data quickly, storing and get access to, is as high as it's ever been. And those fundamental trends are what's driving the demand for our products. And they're very strong. Now the reality is how our immediate customers place their demands is a little spiky. So last year was a phenomenal year for our data center business. We grew 21 percent last year, which was roughly 2 X greater than what we thought when we entered the year, and now this year we've said we'll be down mid single digits. I draw a line to those two in terms of the real steady end demand for our compute, for our storage, and for our movement of data. Those end markets are very strong, and that gives us real good confidence in the outlook for our collection of data centric businesses.
Arjun: So as you look over the next few years, is Intel still very much committed to China?
Bob: Oh sure. Yeah it's a big market for us and in a lot of ways as I indicated, and we have strong customer relationships, and we hope and expect that we can continue those.
Arjun: One company that's really been dragged into this trade fight of course is Huawei, the world's largest telecoms equipment maker, one of the large world's largest smartphone makers as well, they've been put on the US blacklist and that ultimately restricted some U.S. companies from selling to Huawei. There was a reprieve. Has Intel been able to get licenses now to sell to Huawei?
Bob: We're in the process of it. So when the entity list originally came out I think for the most part the whole industry, until they could determine what the entity list really meant, we all stopped, because for Intel we obviously want to continue to serve our customers, but we have to do it while operating within the rules of the countries in which we operate. So we pause, determined what we can ship, what we can't ship, and we did continue to ship in the second quarter once we made the determination for the things that were okay, within the rules of the law. But the process now is that we file licenses to try to get approval. And we've submitted quite a few licenses for our product and the administration had announced, the U.S. administration had announced last week that they're going to try to process the licenses from the industry in an expedient fashion, so we're encouraged by that.
Arjun: And could you give a little bit more detail about precisely what products that you have to apply for licenses for? Because when the U.S. administration came out and said that they would allow companies to apply for licenses, it was unclear exactly what they meant, because they talked about products that were not threatening national security, without giving specifics. Could you just give us a bit more insight into that?
Bob: Well for the most part if you look at the lion's share of what we ship into China generally but to Huawei who's a large customer of ours, is it's general purpose compute. So general purpose product that we ship to you know all of other cloud service providers. So that general purpose compute, we don't believe is what should be worrisome if the quest is to protect national security. So I think in the licensing, what the administration is looking for is, to submit licenses - our intention initially was a broad brush - but ultimately we wanted the industry to submit licenses. So we have the opportunity based on what knowledge they have to determine what we can and what we should not ship. So that's what we are in the process.
Arjun: And have you got any indication about timeline, when you may obtain the licenses that you require?
Bob: No, just that, as I said the administration last week came out and said that their intentions would be to handle the license applications in an expedient manner.
Arjun: And the Trump administration is ready to take actions that many see as it's trying to help US tech. But do you think the actions that he's taken against Huawei and China when it comes to the technology front will actually boost the U.S. tech industry and particularly in your industry in the semiconductor space?
Bob: Well I just go back to our belief on global trade, and I talked about the three requirements that we're trying to encourage is, have dialogues, don't restrict trade of general purpose compute, you know, protect IP, and don't use tariffs as a vehicle, because at the end of the day we think that restricts trade. So I mean that's what we're trying to focus on. Meantime obviously we'll deal with the administrative orders that come out. But you know we're really trying to encourage a free global trade environment that we've become accustomed to, because we think it's good for it's good for our industry, it's good for our business because of the relationships that we've built in China over the years and the large customers that we have.
Arjun: I just want to get your take on a few conversations I've been having with people within the industry. And one of those was that, someone I spoke to said they felt there was a growing mistrust of American tech firms as a result of this trade dispute between the US and China. Is that anything that chimes with what you're hearing from your customers?
Bob: We have very strong relationships throughout the Chinese market, and we've built up trust. As I said over the course of 30 plus years I operate in the market. So I don't see that as an impediment for us to continue to serve our customers as effectively as we can.
Arjun: And just and just the second point is that, there is also a growing number of voices who believe that this dispute has been a wakeup call for the Chinese tech industry, particularly on these areas where the U.S. has been so strong - and semiconductors is one of those - and this could actually spur the development of a homegrown semiconductor industry in China, which ultimately would be a negative for the U.S. chip sector. Is that something you see as a threat to your business in the long term?
Bob: Well I think several years ago China indicated a desire to build more self-sufficiency and this is not a new situation. They've indicated a desire to build more self-sufficiency. For us, all that really means is, we need to continue to invest in R&D to build the best products in the world. So whether it's a large cloud service provider, may be trying to build their own Silicon, or whether it's trying to China's desires to be more self-sufficient. No matter what it is, what that means for our business is we need to continue to invest in the R&D so that we deliver the best products in the world and try to convince our customers that we can help them solve their problems rather than having to do it themselves.
Arjun: Bob let's talk about what's in the future for Intel. You've spoken a lot about 5G recently. And when you announced plans to exit the smartphone modem business, you said that Intel will assess other opportunities in 5G. What are those opportunities as you see them?
Bob: Well first I mean, to maybe start a little macro, but the things that we see now in terms of the role that we play in the industry, is the opportunities couldn't be better. So we have a 300 billion dollar TAM which is over 5 x larger than the way we traditionally defined our served market. The opportunity is significant, and that opportunity is a function of what I referred to earlier, which is this insatiable demand for data. And when you create data you need that process it, you need to store it, and you need to move it. And that plays to the technologies that we build. So first the overall market opportunity for our company is big. It's bigger than it's ever been. Secondly, to take advantage of that market opportunity we are evolving the technologies that we're building. So we used to build the CPU that went inside a P.C. or a data center, and now we're building what we call XPUs, which is simply new compute architectures - not just the CPU but the GPU, artificial intelligence, FPGAs. So we expand the role that we play in our customers' lives by deploying more and more technology inside more and more things. Third, for us is continuing to invest in the emerging technologies that are going to continue this flywheel of demand for data. And those are 5G, those are artificial intelligence, and that's autonomous systems. So those three technologies we've been committed to and have invested in quite a bit over the last several years and we'll continue to. For 5G in particular the biggest opportunity for us has always been what we call the cloudification of the network. So we believe in a 5G world that more and more compute will move from the cloud or from the data centers into the network. And that is what we've been investing in. And we see that as significant opportunities. Secondly for the smartphone modem, the criteria we use when we make these investments is, is it a real technology that's going to differentiate growth for the industry? 5G network, the answer is yes. 5G smartphone modem, not so much in terms of driving growth for the industry. The second criteria we use is, is it an opportunity for us to play a bigger role in the success of our customers? For the 5G smartphone modem, we really only had one customer. And the third criteria is, can we actually make money and generate a good return. We have concluded along the way, with only one customer for the 5G smartphone modem, we didn't really see a way to get attractive returns for our investors. So we doubled down on 5G network where we think there's real opportunities. And last week we announced the sale of the 5G smartphone modem to Apple. But we also retained access to the technologies in the event that we need 5G modem for non smartphone applications like APC, or an automobile, where we can meet the three criteria: differentiated technology with attractive growth, play a larger role in our customer success, and generate attractive returns for our shareholders. Those are the three screens we use. 5G is a big investment for us but it's more the network where we think we can meet all three of those criteria.
Arjun: So you don't feel like you've been blocked out of a potential growth area as more and more companies begin to release 5G capable phones?
Bob: No I don't think we've been blocked out of an area that is a real attractive growth prospect that allows us to play a bigger role in our customers' success and generate attractive returns. I think in the in the smartphone arena there's quite a bit of concentration with a few big players and those big players have a tendency to build their own silicon today. So the one left was our customer. We don't really see prospects for decent returns.
Arjun: Bob you announced a partnership with Rakuten precisely around the cloud application of the network. What exactly does that mean for the end customer?
Bob: I mean we're very excited about this partnership in this relationship and in effect what it means is this this network will allow us through a unique innovative architecture to in effect provide services for end customers at lower cost moving through larger pipes with relatively little latency. So it's an added enhanced service at lower costs and allows us to reduce the prices that consumers pay for their service. With that then it allows for more and more innovation to capitalize on the network for other things other than just the data services they buy. But with the 5G at the network edge it creates more applications for innovation for innovation for things like autonomous driving that can benefit from compute happening not just in the car and not just in the cloud but local through low latency high speed big pipes that enhances. We believe the pathway to autonomous driving.
Arjun: Well autonomous driving of course is a technology that can really be underpinned by 5G and artificial intelligence as well. You recently visited Mobileye in Israel had a chance to sit in one of the cars equipped with that that company's technology. What were your impressions? I know Mobileye is on a path to launch this driverless taxi service in 2021. Can you give us some insight into where you see driverless cars going over the next five years and in this particular project for Mobileye as well?
Bob: Yeah. Well first it's we're approaching our two year anniversary since the acquisition of Mobileye we couldn't feel better about them being part of the broader Intel family first so I'll go back six months ago six months ago I was over in Israel and I took the the mobileye drive on the highways of Jerusalem. And it was a flawless experience but was on the highway it's not as challenging as the streets of Jerusalem. And when we did it they said we drove after doing the highway drive. They drove me through the neighborhood. But it was with hands on and they said we're developing the technology so we can drive through the streets of Jerusalem. You should come back and visit in six months. And I was I mean the streets of Jerusalem is its high impact driving. It's intense. So I went back and it was just a few weeks a few weeks ago and I was blown away by how fast and how quickly the technologies developed because we went through the streets of Jerusalem. It was almost a 20 minute drive and not once during that drive. The gentleman behind the wheel have to take his hands and put them on the steering wheel. Their technology navigated the streets of Jerusalem which increasingly builds our confidence for our ability to go from L2 to L2 plus to L3 and to deploy automobiles for a robo taxi with our partner Volkswagen and champion motors in early 2022. So the technology continues to advance the opportunity we set. The opportunity set we see with the technology is even greater today than our lofty ambitions when we made the acquisition two years ago.
Arjun: And so when do you expect driverless cars to become widespread?
Bob: Well I think first our belief is that applications like robo taxi will happen before broad based adoption of L5 autonomous driving. And the reason is really twofold. One we believe the acceleration of autonomous driving will be greatly enhanced by the deployment of 5G and that's going to take a few more years to be deployed. And secondly and maybe more importantly in the absence of compute at the network the cloud is a little too far away. And therefore you have to put more and more compute in the vehicle itself and put more and more compute in the vehicle itself to manage the intricacies of autonomous driving can be pretty expensive and to get broad based adoption in vehicles for autonomous driving. You can't have extremely expensive compute in the vehicle. In a robo taxi environment you can because you get to monetize that vehicle 24 hours a day seven days a week and therefore you can afford to put more compute in the vehicle. Our belief is robo taxis will create more adoption first and then through cloudification of the network and the access to that compute technology for vehicles on the road will come after that. I'm really excited by the opportunity.
Arjun: And potentially I mean a key part of this is around A.I. chips as well. They play a key role in many different areas. How will Intel play in the A.I. chip space because I know you announced a partnership with Facebook to try and develop an A.I. chip for later in 2019. So is that kind of the path you're going to go down partnering with companies to create specific solutions or will you have a general purpose A.I. chip as well. What is Intel's strategy?
Bob: Well I mean first strategically our objective is to build A.I. functionality in everything that we and everything we design and make. So what that means is even for the Xeon general purpose compute we've built significant A.I. functionality including a recent technology that we've referred to as DL boost that enhances the performance of the general purpose compute by building A.I. into that in and of itself. But in addition to that we also believe that there's workloads and used cases where a separate chip is really relevant. So we have the Facebook example that you referenced is a discrete A.I. inference chip. And then even recently we announced it with Baidu in China a discrete A.I. training chip as well. So strategically for us it's how do you build a variety of different compute architectures where A.I. needs to be incorporated into all the architectures. But also you have to supplement that with A.I. chips specifically both for inference and for training and that's been a really big high growth market for us over the last couple of years. And we expect it to continue going forward.
Arjun: Bob let's talk about your leadership of Intel. You've been in the job for a few months now. A key part of your strategy is about diversification. And so when we look at Intel in the next five years how do you expect the revenue mix to change?
Bob: Well you know it kind of starts with this transformation that we've embarked upon to go from a P.C. centric to a data centric company to a company that builds the technologies that powers the world. What that means is a relative five or so years ago we were predominately a P.C. centric company where over 70 percent of our revenues came from a CPU inside of the P.C. Last year, with our investments in this more data centric world we were more 50/50. So fairly dramatic shift in a relatively short period of time to go from 70 30 P.C. centric to 50 50 P.C. and data centric as we go forward the size of the market and the growth rate of the data centric market and the collection of businesses and technologies that we have to deploy into that environment, we expect that to be a bigger and bigger chunk of the company going forward. So if it was 70 30 five years ago roughly 50 50 today at some point in the future I would expect it to be more like 30 70. With the technologies and the architectures that we're building for this data center of the world will become a bigger and bigger part of the company.
Arjun: And as you go through this transformation there are challenges as well there's not just the political environment that challenges say memory pricing. Of course Intel has been accused as well by some analysts of falling behind on some key technologies like 7 nanometer chips. How do you expect to navigate the company through this tricky time?
Bob: Well I mean first it starts with one hundred and eight thousand of the best people we believe in the industry ensuring that those hundred and eight thousand employees are rowing all in the same direction to deliver for our customers. So that's been the recipe that's worked for us very well over time. We expect to continue that. In terms of process technology, the real challenge for us now is to get systems on shelf for the holiday season this year for our 10 nanometer which would be the equivalent of competitor seven nanometres to get our 10 nanometer systems on shelf in the fourth quarter. We announced just last quarter that we've qualified the product we've started shipping to customers and our yields are improving. So we feel good about where we are on 10 we also announced that 7 nanometer which should be the equivalent of competitors 5. We believe we'll be back on a two year cadence. So back on that two year moore's law. Two times density within two years we'll launch 7 nanometer products in 2021. So our focus and intensity on process technology now is to ramp 10 and get to 7 back in a two year cadence. But more importantly than simply the process technology today what our customers are looking for is product leadership. Process technology is important but it's not the only thing. That's process technology and packaging technology. It's not just CPU architecture but it's GPU A.I. and FPGA architecture it includes the role of memory. Memory becomes an increasing component in our customer's bill of materials and it needs to keep pace the innovation and memory needs to keep pace with the increased processing capabilities. Security is extremely important interconnect and then how do we take software or one API layer for our fifteen thousand strong software developers and the external developers that build software to make the performance of our products better. How do we make it easier for them across a multitude of architectures so that with process technology and product technology coupled with software that we're delivering product leadership for our customers going forward. So process technology is deliver 10 get to 7 quickly but couple that with the other technologies that are one hundred eight thousand strong employee base are delivering for our customers.
Arjun: And 2021 seems to be this year that you feel Intel will be back in in the leadership position. I saw a quote actually from you that said until then though Intel could be in a bit of a bathtub. What makes you confident that by that time Intel will get its groove back?
Bob: Well first I would say just to put it in the context 2018 was our 50th anniversary and it was the best year in the company's history. So yeah and we grew you know kind of double digit rates of growth in our last three years we've added. Twelve billion dollars in revenue with relatively little cost. So we feel very good about where we are and our expectations are to continue to play a bigger role in our customer success. I think as we said so that we have very good momentum and we are playing a much bigger and broader role for our customers as we go through it from 10 to 7 nanometer at the same time. We're gonna be developing two process nodes so the implications of developing two process nodes at the same time. The good news is we're advancing the technology. The challenge is when you do that you have a tendency to have a bathtub in gross margins not in demand for your product per se but a bathtub in bringing on two new process technologies in roughly the same period of time has a tendency to compress gross margins. But those are great investments to make in over the medium to long term they're well worth it.
Arjun: Bob thank you so much for joining me today. I really appreciate it.
Bob: Thanks a lot I really enjoyed it.
END
Media Contact:
Clarence Chen
Communications Manager APAC, CNBC International
D: +65 6326 1123
M: +65 9852 8630
clarence.chen@cnbc.com
About CNBC:
CNBC is the leading global broadcaster of live business and financial news and information, reporting directly from the major financial markets around the globe with regional headquarters Singapore, Abu Dhabi, London, and New York. The TV channel is available in more than 415 million homes worldwide.
CNBC.com is the preeminent financial news source on the web, featuring an unprecedented amount of video, real-time market analysis, web-exclusive live video and analytical financial tools.
CNBC is a division of NBCUniversal. For more information, visit www.cnbc.com