切換側邊選單 切換搜尋選單


Everybody’s Partner


Everybody’s Partner


Fabless semiconductor giant Qualcomm has a hand in a third of Taiwan’s high-tech exports. In an exclusive interview, CEO Paul Jacobs discusses the company’s business model, its partnership strategy and its relationship with Taiwan.

Everybody’s Partner

By Hsiao-Wen Wang
From CommonWealth Magazine (vol. 521 )

No company can be an island. That realization has made a deeper impression on 49-year-old Paul Jacobs than any other nugget of wisdom since he became the CEO of Qualcomm Inc. in 2005.

Qualcomm has many faces – the world’s biggest mobile phone chip supplier, TSMC’s biggest customer, HTC Corp.’s closest ally, and MediaTek’s strongest rival. Its global reach is unparalleled, with a hand in a third of Taiwan’s tech exports.

Qualcomm’s market value has surpassed that of American chip-making giant Intel Corp., and Jacobs has replaced Microsoft’s Bill Gates as the featured speaker at the opening of the annual Consumer Electronics Show in Las Vegas. But he still frowns at being seen as the industry’s hegemon, preferring the designation “partner.”

That’s because creating a high-tech platform and managing partnerships is Qualcomm’s secret to success.

At the company’s headquarters in San Diego, the board of directors’ meeting room on the 10th floor features a 15-meter long white marble table, as smooth and shiny as power itself. The boundless blue sky and green valley seen from the office’s floor-to-ceiling window resembles the clarity of Qualcomm’s vision.

Jacobs, who has also been Qualcomm’s chairman since 2009, shuns sitting at the head of the table, preferring instead a seat in the middle to listen to the ideas of others.

“He is a very sincere person,” says TSMC CEO and chairman Morris Chang.

Inside the brown-haired brown-eyed Jacobs seems to live a young boy – his laugh carries a trace of innocence – and his eyes radiate the passion of an engineer. When he first walked into a TSMC foundry, he could barely contain his excitement.

“I remember the first time I went into (TSMC’s) fab, and it’s like Star Wars, all these little bots going across the roof, and all the stuff,” Jacobs recalls.

In 2012, Qualcomm was unable to meet its CPU supply commitments, leaving Jacobs feeling deeply apologetic to his friend HTC CEO Peter Chou. The imperatives of business have more than once taken an emotional toll.

But Jacobs has another side, one featuring a steely will.

Facing the rise of cheap, non-branded mobile phones in China, Jacobs decided to attack this low-end market, ready to sacrifice the company’s margins.

Sacrificing Margins

The move took aim at Taiwan-based MediaTek Inc., the dominant player in supplying China’s “bandit” phones with essential chips.

“800 million ‘made in China’ mobile phones are sold to China and emerging markets a year. Even as competitors have entered this market, MediaTek’s market share has continued to rise,” says MediaTek chief financial officer David Ku, brushing aside Qualcomm’s incursion into MediaTek’s stronghold.

Internet scribes often like to characterize the contest between Qualcomm and MediaTek as a “battle between the gentry and the proletariat,” but Qualcomm may not be in a stronger position.

“Qualcomm has nowhere to go when it comes to selling chips,” says Barclays Capital Securities chief analyst for Asia Andrew Lu. Qualcomm has 300 engineers supporting customers in China, compared with MediaTek’s 1,000, and it offers customers 100 reference chipset designs, compared with more than 1,000 from MediaTek, Lu says.

“ZTE mobile phones that use Qualcomm’s quad-core chip processors sell for 1,234 renminbi. (Phones that) use MediaTek’s quad-core chips sell for 1,800 renminbi,” Lu says, a sign that MediaTek is moving up the value chain in the battle for China’s market while Quadcomm is sliding down.

But Qualcomm’s profit structure remains stronger than MediaTek’s, mainly because technology licensing accounts for 70 percent of the company’s earnings. For every handset sold around the world, Qualcomm earns US$5 to US$7 in royalties, and that includes handsets made by MediaTek’s customers.

This gives Qualcomm a steady cash flow accounting for 30 percent of its business revenues, a gravy train that enables the company to invest 20 percent of its revenues every year in R&D and explore next generation telecommunication technologies.

The broad group of allies, technology licensing and R&D investment all contribute to the positive cycle seen in Qualcomm’s operations – a cycle yet to be successfully imitated and one that inspires envy among low-margin high-tech operators in Taiwan.

Qualcomm CEO Paul Jacobs recently joined CommonWealth Magazine in a discussion of the high-tech industry, the role Taiwan plays, and why the horizontal model has met with grand success. Following are highlights from the interview.

Q: How did Qualcomm develop its business model back in 1985? How did you figure out you could build a licensing business when you had only 50 patents?

A: That’s a funny story, because it didn’t start as it is now. Back then we needed money, just to keep the company going. So the licensing business was originally about upfront fees. And now we have the ongoing licensing almost as an afterthought.

Q: That’s it?

A: Yes, that’s how it happened, seriously. It didn’t happen like many things in companies. There’s a sense of preparation. And there’s sense of opportunity. Put these two together, and you get some luck, right? So, it was that.

You know, sometimes you have to do things, and they turn into something good. We started out thinking we have to be vertically integrated. We had handsets, we had infrastructure, we invested in operators, we had the chipset, we had the technology. I was running the handset business at the time, and we were working so hard, and then Samsung came in as a competitor. They started out as our close partner, we were working with them getting the technology out there, and then relative to my business, they were my competitor. We worked really hard, and we made three phones throughout that year, and Samsung would come into our wireless operators where we showed our three phones and would say, “Here is what we have. Here’s our 20 phones. Choose the one you want.”

Q: They already had the scale?

A: Yeah, they scaled much better than we did. They had the experience, because they had been building consumer products. We said, you know what, why are we doing this? We had a few problems, quality problems, things that cost money. Why are we doing this? They are better than we are. Let’s get out of the handset business and let them do that. We can be the best partner we can be to them. And we will go focus on what we are good at, which is creating the new technology and designing the chipset.

Q: You started as a vertically integrated company, but now you choose a horizontal business model. What’s the advantage of that? What’s the power of that?

A: The power is that we can focus all our efforts on the things we are good at, which is: what’s the next technology that will enable other people to do something incredible for the consumers? We never served the consumers well. That just wasn’t our skill set at the time. So having partners like Samsung, like HTC…

Q: You turned enemies into partners?

A: Yeah, they are very close partners now. We work hand in hand. Even today, you have Samsung doing some of their internal development – RF, AP. It’s okay, though. It’s friendly competition. And then I’ve known Peter Chou from HTC for so long. He and I just trust each other. He and I actually sit down and brainstorm on things.

One of the big things for me is learning to partner well.

When I had my first all-head meeting, all-company meeting when I took over (in 2005), I got up, and I said, we were known for innovation, and we were known for execution. But we were not necessarily known as good partners at the time. People thought we were too arrogant, we didn’t support them enough, we were trying to get our things instead of theirs. And I said, no, our three values are innovation, execution, and partnership. And everyone has taken that to heart in this company. And that’s what enables the horizontal business model. And now we are partnering with everybody. I have people coming to me and saying, how can you be the partner of these two companies when they hate each other?

Q: Like Samsung, Apple, HTC?

A: Yeah. But we focus on going to that company, and saying, what do you need? What can we do to help you?

And you know why HTC is such a good partner for Qualcomm? Because HTC, Peter (Chou) and Cher (Wang), they always want to take the newest thing and take it to market as fast as they can. For a company like Qualcomm, built on technology and innovation, having a partner that can take technology as fast as possible into the market is really important, because that’s pushing everybody else in the market. So you want to find people who are great leaders. That’s why the horizontal model works.

Q: Why is the “fabless plus foundry” model more powerful than IDM (integrated device manufacturer), especially in the age of mobile Internet?

A: That also started early on. We couldn’t afford fabs. So the funny thing is that we actually fabbed with Intel in the very early days of Qualcomm. Then TSMC came along and made it a business model. For us, it was important. If you own a fab, you spend time worrying about filling the fab, like utilization and yield rate. We don’t think about that that much. We think about that only when there’s a shortage. And we said, hey, get more machines and build it up faster. But generally, that’s not how we spend our time. If you own a fab, you are thinking all the time, “Am I using my equipment to full utilization?” We spend our time thinking about innovation. I spend almost no time on supply chain issues. Last year maybe I spent a little time. But mostly, I don’t. And I think that helps us a lot. Most of our managers spend more time on innovation than on supply chain issues. And therefore we can stay ahead of the competition.

Q: Why do you think proponents of the IDM model, like Intel, could not comprehend the power of the “fabless plus foundry” model?

A: I think because they felt that was their core strength. By investing in the manufacturing and process technology, as long as they could fill up their fabs, they believed they would have a competitive advantage. And problems happened when their transistor advantage didn’t translate into system advantage, because their system wasn’t as good. Our ecosystem, and the ARM ecosystem, we build our specific microprocessor within ARM’s instructions. We can optimize the entire system for low power. They might have the transistor advantage right now, but that’s going to change over time. And certainly TSMC is investing very, very heavily to make sure their transistor roadmap is equal to Intel’s.

Q: To be more specific, is TSMC’s 28 nanometer HPM process technology better than Intel’s 14 nanometer finFET? Intel always says that they are one or two generations ahead.

A: Right now we probably have a transistor advantage, or we’re even. They are going to have a transistor advantage when they go 14, or whatever they call it. They might have a transistor advantage. But they have had a transistor advantage over us in the past. But we have the system advantage. And then we are going to the point of transistor parity or advantage for the next couple of generations, so then the system advantage we have will be much better.

Q: Why do you think you will have a transistor advantage over Intel in the coming years?

A: Certainly when they reach 14 or 16 nodes, whatever the number is, I think we will be roughly at parity. Not much time difference. And then at 10, I think there will be process and performance parity.

Q: Going back to your business model, could you give some examples of how you reduce cost and time-to-market to encourage new market entrants?

A: Right now we are really focused on that, because at the low end of the market, you see a lot of new entrants. They are small, they don’t have a lot of engineering resources, and you see what happened in the GSM space, especially in China. A lot of these small companies took huge, huge market shares. So you have these Qualcomm Reference Design programs – companies are really, really fast taking our stuff to market. One of the bigger ones, Tianyu – they are one of our customers for the latest QRD designs. They took our silicon-on-dot to carrier acceptance in 60 days. We have smaller companies, like Chili – they just took our reference design, and they got it out in 60 days. So by giving our customers more and more solutions, they are able to get to the market quickly. It depends on the partners. Some partners have a manufacturing advantage, some have a brand that is known, some have a distribution advantage, some of these companies are building their own cloud services. There’s really a lot of small niches.

So our partner strategy at the low end is to help that. Of course, on the high-end, HTC is a great example. HTC One just came out. They took our cutting-edge Snapdragon 600 chipset, and boom, brought it out quickly. They are the first one out. We are really close. Our engineering teams are there working hand in hand with them in Taipei or they are here.

Q: When you reach down to the low end, how will that squeeze your profitability?

A: I think at the low end we are going to fight it out.

It can be lower margins. That will happen for a while. I think it’s just going to be a battle down there. For us, it’s good. We want to see more people using smartphones. And you have to get the price very low to get to the mass market. That price war on the chipset helps us grow our overall business. And people will buy it up. So our strategy is to fight it out at the lowest end and have a set of products above that.

Q: You have a complete portfolio?

A: Yeah, and that’s how we are going to fight that battle. And at the low end of the market, sure, it’s going to be whatever it is. We are going to fight that battle. Because I don’t want to give opportunities to competitors to build a big base down there and eat their way up.

Q: You are taking lessons from TI and Nokia…

A: Oh, yeah, Nokia in particular. In GSM they really got beaten up by those little guys eating up from the bottom. And we saw that. We watched that really carefully. We prepared for that attack. So we are ready. We have really good products for that market. We will be very competitive down there.

Q: What’s the biggest lesson you’ve learned after taking over as CEO eight years ago?

A: Let me put it this way: We were sitting in San Diego and thinking that everybody was appreciating what we are doing. In fact, there were companies that looked badly upon us. I took over, and within months we were sued by Nokia, Ericsson, Panasonic, NEC, Broadcomm, TI, the Korean government, the Japanese government... I mean, you wake up and realize, boy, there are people who really don’t like us. So I really learned that being a good partner is really the most important thing. Don’t ignore people. Make sure you touch all parts of the ecosystem, so that they see us as an enabler.

Q: An enabler instead of an exploiter?

A: Yeah, the story that was told is that our royalty business was tax collection.

Q: Is the royalty business bigger than your technology department?

A: No, but we spend US$4 billion dollars on research and development, and basically that’s for our partners. It’s really important for me to make sure that we tell the story of what we are doing to help people. But more than just talking about it, we do it. When our partner has an issue, our engineers show up. They have experience. They solve the issues, they help people out. Therefore, people see that the money they paid for the royalty means that they got the technology and the support to make that into reality. That to me is the biggest lesson. To be a good partner, to choose consciously not to compete with your partners, but to enable your partners. That really works well.

Q: What about (TSMC chairman) Morris Chang impresses you most?

A: I think it’s his focus. It is a very difficult job he is doing. I remember the first time I went into his fab, and it’s like Star Wars, all these little bots going across the roof, and all the stuff. What’s really important about that is, in a traditional fab factory, there are lines of machines, this machine goes to that machine. That’s not what TSMC is at all. TSMC is like: machine here, there, here, there. To think of the IT system, to handle that wafer capacity, to build the process technology. When Intel said they are going into the foundry business, that’s not their skill set. That’s not how their factory is laid out. Intel’s factory is static and repetitive. It’s the opposite. Intel is no flexibility, whereas TSMC is infinite flexibility. Infinite is probably too strong a word, but you know what I mean.

He has the vision, and he executes it to the point where we are at cutting-edge technology. This is not an easy business.

And he wants to invest in the future, like us. We put a huge amount of money into our R&D and create new chips long before we have customers for them. We put huge amounts of money into wireless technology before we know that wireless operators will even buy it. And sometimes they don’t, and we waste a lot of money.

And he is the same way. He invests a huge amount of money building a fab before he knows who the customers will be.

Q: What do you hope to achieve with Taiwan’s tech industry?

A: Between 2012 and 2016 there will be another 5 billion smartphones, more than half going to emerging markets. So being able to address both the high-end and low-end smartphone market is important. That’s what the current smartphone business is like.

The other thing, especially with TSMC, is to drive the roadmap. Intel obviously wants to come into the mobile space, and we need to make sure that we get the process technology so that we can be competitive on the transistor level and beat them on the system level. That’s very important.

Then of course you have the whole computing business. Computing is completely changing. And we look beyond that, we see the Internet of Everything, the phones, the tablets, sitting in the middle of the world where everything is connected. And obviously, there are people like Foxconn, who make every kind of gadget, anything you can think of. There’s huge opportunities there.

Qualcomm and Taiwanese companies are really related throughout the whole value chain. I really like it, all the way from the chip to finished systems. It’s been a very good partner for Qualcomm.

(Interview compiled by Hsiao-Wen Wang)

Translated from the Chinese by Luke Sabatier


Fiscal 2012 (year ended Sept. 30, 2012) Sales: US$19.12 billion, up 28%

Fiscal 2012 Income: US$6.46 billion, up 20%