Why Intel will continue to shrink even in the absolute best-case scenario

“I don’t want to see you in my office again until you have a marketing plan that includes stunt pogo, and that’s final.”
“I don’t want to see you in my office again until you have a marketing plan that includes stunt pogo, and that’s final.”
Image: AP/Invision/Jon Furniss
We may earn a commission from links on this page.

For years, Intel has been shrinking. Quarter after quarter (after quarter), the company’s performance has roughly tracked the dismal showing of its biggest customer, the PC industry. In the press release accompanying Tuesday’s disappointing quarterly results, Intel’s newish CEO Brian Krzanich, for whom this is the first full quarter at the helm, sounded properly humbled. “In my first two months as CEO, I have listened to a wide variety of views about Intel and our industry from customers, employees and my leadership team.”

But Intel’s chief also took pains to highlight the products that are supposedly going to save the company, or at least halt its slide: the company’s Atom processors, which can be used in smartphones and tablets. “I’ve made it Intel’s highest priority to create the best products for the fast growing ultra-mobile market segment,” writes Krzanich. In other words: Intel’s new chief is prioritizing the microchips that are powering the devices that are killing the PC industry—an industry on which Intel has long depended.

Why Intel can’t win, even if it wins

Here’s the problem: Even if Intel were to win a significant victory in the battle for market share against the many, many other companies that make chips that power mobile devices, back of the envelope math suggests Intel still couldn’t make up for the shortfall in revenue that is resulting from declining sales of PCs. (Keep in mind that such a victory is also totally unrealistic, given that Intel currently has less than 1% of the global market for processors for mobile devices.)

Update: Chris Kraeuter, spokesperson for Intel, disagrees. Essentially, says Kraeuter, Intel’s potential addressable market is much bigger than just smartphones, and includes the still-nascent (when compared to smartphones) market for tablets, as well as “embedded systems,” which includes things like set-top boxes, smart TVs, in-dash systems in cars and countless other devices. “Any product that’s considered a computing device we will target,” says Kraeuter. New CEO Krzanich has started an internal crusade to realize this vision, pushing the company’s engineering and manufacturing arms to get Intel’s most cutting edge technology into the company’s mobile chips as quickly as possible, so that Intel can offer a level of power and performance that can’t be matched by competitors like Qualcomm and Samsung.

In the short-term, at least, analysts aren’t buying this reasoning. As Michael McConnell of Pacific Crest Securities notes, for every mobile chip Intel sells, it makes as little as $20. For every PC chip Intel sells, it makes about $100. The margins on the chips that power mobile devices and PCs are the same, but the chip in a PC sells for about five times as much.

Which means that for every PC that fails to sell because of a tablet, smartphone, or the missteps of Microsoft, Intel has to get its chips into as many as five times the number of mobile devices. In 2012, approximately two-and-a-half times as many mobile devices as PCs were purchased (800 million compared to 300 million), and PC sales are down 14% year-over-year, and analysts have declared all of 2013 a write-off.

Intel has proposed that the ever-expanding infrastructure of the cloud required to feed data to all those mobile devices will also demand a substantial number of Intel chips, but there are a variety of reasons the cloud will hardly be Intel’s savior.

So even in a scenario far more optimistic than any analyst could credibly advance, in which Intel quickly makes huge advances in the market for mobile devices, the company’s revenue will continue to shrink. And there are signs Intel has made these projections, itself: Having originally projected that it would spend $13 billion on research and development in 2013, the company has since revised that figure to $11 billion. That’s not good news for a company that cannot survive unless it continues to stay at the forefront of the difficult and highly competitive race to manufacture microchips.

The root of Intel’s problem

There are a number of reasons for the difference in cost between mobile and PC chips, and one is that the chips that power mobile devices are simply smaller, simpler, less technologically advanced and cheaper to manufacture than the high-end PC chips Intel sells. Related to this, there is so much more competition among manufacturers of these chips. In PC chips, Intel is fighting a single ailing rival, AMD, and both companies sell chips that speak a language (known as an instruction set) that was originally defined by Intel. In mobile chips, the lingua franca has been defined by ARM Holdings, which licenses its instruction set and/or chip designs to anyone who wants them. This means that in mobile chips, Intel is fighting a multi-headed hydra consisting of all of the ARM licensees in the world, including Samsung, Apple, Qualcomm, NVIDIA, and countless ultra low-cost Chinese chip makers like All Winner, which churns out the brains of the world’s cheapest tablets.

It has to give Intel executives night terrors to know that for less than the profit Intel makes on a single PC chip, people in developing markets—in other words, the “next billion” consumers who are driving growth for so many other multinationals—can buy a tablet that, for many of them, is a more-than-good-enough replacement for a PC. Computing is now cheap, ubiquitous, and democratized in a way that’s great for humanity, but a problem for Intel and its shareholders. It’s certainly not the case that Intel is going to go away, but it’s also not clear when its declining revenues will hit bottom. To answer that question, analysts are going to have to start evaluating the potential for Intel’s next generation of Atom processors to become as ubiquitous as competing chips from ARM—and certainly, it doesn’t look like a turnaround is in sight until 2014.