Qualcomm's Biggest Problem: Meeting Demand

Image: Getty Images

Qualcomm CEO Paul Jacobs: Sailing through choppy seas

Even in the middle  of the fray, Qualcomm CEO Paul Jacobs somehow seems to float above it all.

Back in April Qualcomm warned that it is struggling to meet demand for its cutting-edge mobile phone radio chips. The news has troubled both shareholders—the stock has since sold off 15 percent—and customers, who aren’t able to make as many high-end smartphones as they might otherwise like. Meanwhile, the company is seeing intensifying competition from, among others, Samsung, which also happens to be a key customer. And neither Jacobs nor any of his minions will talk about the company’s marquee customer, Apple.

And yet, amid a roomful of re-porters at Qualcomm’s developer con- ference at Hilton San Diego Bayfront in June, the 49-year-old engineer seemed immune to the drama.

Will the company take on Intel in server chips? “It’s certainly an area we are looking at,” Jacobs said. Will Qualcomm build its own factories to keep up with demand rather than continuing to farm out production to contract manufacturers like Taiwan Semiconductor? “Our inclination is to retain the model we have,” Jacobs said. Will Qual- comm build its own gadgets? “No.”

Then Jacobs paused to glance out the Hilton’s enormous fourth-floor window at the glittering blue waters of the harbour below. “Wow,” Jacobs said. “That’s an enormous boat.”

There are reasons Jacobs can main- tain his Zenlike calm: Despite all the concerns, business is good. Qualcomm will generate $15 billion in sales for the fiscal year ending in September, up 36 percent from $10.9 billion during fiscal 2011, analysts estimate. Qualcomm’s chips can be found in leading smart-phones from Samsung, HTC and Nokia. Every Apple iPhone 4S sold contains $23.54 worth of Qualcomm parts. Better yet, Qualcomm collects royalties on nearly every smartphone sold. And the company does all this while steering clear of the messy business of building its own handsets or running its own chip fabs.

That’s a profitable bit of triangulation: Qualcomm’s gross margin of more than 60 percent is richer than any of its major customers’, including Apple’s.
That hands-off model, however, hit a bump earlier this year. Tight capacity at its manufacturing partner, TSMC, forced it to offer disappointing earnings expectations. More ominously, one of its most successful customers, Apple, uses Qualcomm’s modem chips but skips the company’s ARM-based microprocessor designs in favour of its own, keeping more profits for itself. How long can Qualcomm continue to produce growth rates so far ahead of the rest of the chip industry? Should it build its own fabs? And can the company resist the temptation to re-enter the handset-hardware business?

For Jacobs, those aren’t abstract questions. Under his father, Qualcomm co-founder Irwin Jacobs, Paul ran Qualcomm’s handset business. It was one of Qualcomm’s biggest flops, racking up $200 million in operating losses before it was sold to Japan’s Kyocera in 1999. “One of the reasons was that our customers felt like we were competing with them,” Jacobs says in an interview with Forbes. And without those customers Qualcomm would never have become big enough to invest in better wireless chips. “I think it’s just a  scale issue. Yeah, if you have a lot of scale, then you can invest, but ‘Can you invest enough?’ is the question.”

Thanks to Paul’s father, Irwin—his predecessor—the answer to that ques- tion is yes. Under Irwin Jacobs Qual- comm was best known for pushing a protocol known as CDMA, for “code division multiple access”, that allowed carriers to cram more calls into the same amount of radio spectrum. A version of that technology—dubbed WCDMA—would become the basis for the third-generation wireless technologies now built into most of the world’s smartphones, unleashing a geyser of licensing revenues.

That gush of cash—about a third of Qualcomm’s revenues now come from licensing—funds a research and development juggernaut. In the fiscal year ended September 2011 Qualcomm spent $3 billion, or 20 percent of its revenues, on R&D. Some of those efforts can get pretty wild, such as “augmented reality” software that meshes mobile camera images with virtual reality. Other researchers are working on sensors that will sit in a patient’s bloodstream and lodge in his wrist.

“Imagine your phone rings and says, ‘Go to the doctor now, because you need to manage your heart,’ ” Jacobs says. “You’ll be a mobile phone, too.”

The wild ideas don’t always work. Among the flops: Zeebo, a gaming console for the developing world Qualcomm helped develop, and the SmartBook, cheap notebook computers powered by smartphone processors. “They throw a lot of things against the wall and see if they stick,” says Will Strauss, president of wireless market research firm Forward Concepts.

That hit-or-miss strategy has generated two major hits since Paul Jacobs took over as chief executive in 2005. In 2007 Qualcomm launched its first Gobi chips, multiband radios that can be configured to support networks used by different carriers. The first chips were bulky—and confined to laptops—but Qualcomm shrank the chips and added new capabilities.

That kind of dexterity is paying off: Qualcomm grabbed 45 percent of the $15.1 billion market for wireless modem chips last year. Thanks in large part to their ability to handle the growing welter of competing networks used by wireless carriers, Qualcomm’s wireless modems are now built into all Apple’s iPhones and iPads. ‘They can ship the same chip, and it can be used with Verizon or AT&T,” says Strauss. “They’re the only company in the world that can do that right now.” Jacobs’ most aggressive move has been into the processor business. Introduced in 2006, Qualcomm’s Snapdragon chips are based on the same ARM-instruction set used by most mobile phones. Its chips, however, include Qualcomm’s multimode wireless modems, allowing phones using the chips to use less power—and take up less space—than chips that include separate processors and modems.

Because Qualcomm’s wireless modems can now handle a broader range of protocols than competing chips, the Snapdragon processors have been built into a wave of high-end smartphones from Samsung, Nokia and HTC, among others. Qualcomm grabbed 50 percent of the $7.9 billion smartphone processor market in 2011.

Qualcomm’s biggest problem now: Meeting demand. That’s a side effect of the company’s “fabless” strategy, relying on contractors to make the stuff Qualcomm designs and sells. The recent issue was a lack of capacity to produce cutting-edge parts with features just 28 nanometers wide, targeted at high-end smartphones.

That’s a problem that rivals Intel and Samsung will never have; they make their own parts. Yet despite Qualcomm’s size—it’s now the biggest chip company that doesn’t own its own fabs—Jacobs has no plans to build or buy a factory. “Your costs get set by how full you can make the factory,” Jacobs says. “What happens, when you run a fab, is you end up worrying all of the time about, ‘What am I doing to fill the fab?’ ”

Qualcomm’s manufacturing hic- cups, however, give rivals a chance to catch up. While Intel, which is promising to offer a chip able to support the latest LTE networks by year-end, could be Qualcomm’s highest-profile foe, smaller players like Nvidia and Renesas Electronics could pose bigger problems, analysts say. Renesas has chips that can compete, on paper at least, but they’ve been missing in action in the US market. Analysts see it as a possible acquisition target.

Nvidia could be a bigger headache. With the acquisition of wireless chip company Icera last year, graphics chip designer Nvidia will introduce do-it-all chips that can support both high-speed LTE networks and slower third- and second-generation networks early next year.

For now Qualcomm’s weakest spot could be tablets. Led by Apple’s iPad, Credit Suisse estimates sales of tablet computers will grow to 169.7 million units in 2013, up from 110 million units this year.

And while Qualcomm’s wireless modems are an option on Apple’s iPad, the device’s silicon brain remains Apple’s own A5X processors, an ARM-based design comparable to Qualcomm’s Snapdragon chips built for Apple in factories owned by Qualcomm customer Samsung.

Qualcomm will get a chance to get more of its parts into tablets later this year with the launch of Windows RT, a version of Microsoft’s Windows 8 operating system that can run on ARM-based smartphone processors such as Qualcomm’s Snapdragon. So far, Nvidia seems to have an edge here. Microsoft unveiled Surface in June, a pair of tablets built around mobile processors supplied by Intel and Nvidia. Google’s Nexus 7, also introduced in June, sports a 7-inch screen and a $199 price tag, less than half the iPad’s $499 starting price. It also uses Nvidia’s Tegra 3 processor.

Jacobs says a tablet built around its chips will arrive with the Windows 8 launch in October. The tablet will reportedly be built by Samsung, although Jacobs declined to name his partner.

“I think you’ll like the products when they come out,” Jacobs says. If it’s Samsung, that means one of Qualcomm’s biggest competitors will be forking over cash for one of Qualcomm’s chips, again. We should all have such complications.