Paul Otellini, Intel's outgoing CEO and a 40-year veteran of the company, gave his last earnings call last week. And the results were, well, a little bit less than sanguine. The company’s profits in the first quarter of this year were US $2 billion, down 25 percent from the last quarter of 2012.
That's a pretty big decline for the world's biggest chip company. And the culprit, if you had to choose just one, seems to be declining PC sales. A few weeks ago, analysis firm IDC reported that the number of PC shipments in the first quarter of 2013 was down 13.9 percent from what it was a year ago—the steepest such decline since the company started tracking shipments in 1994. Late last week, private equity firm Blackstone said it is dropping plans to cobble together more than $24 billion to buy Dell, citing the health of the PC market as one of the reasons for the change of heart.
Intel, in theory, is in a much better position to rebound from declining PC sales than Dell is. Consumers are spending their PC money on new tablets and smartphones, and Intel has been gearing up to make a dent in the mobile market. This year, it is expected to begin releasing a series of Atom chips that boast a microarchitecture, code-named Silvermont, that is truly optimized for low power. The campaign will reportedly start with chips for low-power servers, followed by tablets, and then smartphones. (For those who want to follow along, the code names for those chips are Avoton, Bay Trail, and Merrifield, respectively.)
How much of an impact these chips will make is anyone's guess. Although Intel has already inked a few deals with mobile companies, this is a big departure for the company. "The PC is like a huge flywheel driving you in one direction," analyst G. Dan Hutcheson, CEO of VLSI Research, told me last year, not long after Otellini announced his resignation in November. The challenge for the company will be whether it can "become as light and as fast-moving as Qualcomm and Nvidia."
There is another big question mark in Intel's future: the company's direction after Otellini leaves in May. A replacement CEO has yet to be named, but that hasn't stopped a lot of speculation about whether Intel will take a drastically different tack. One idea that has been circulating for years is that Intel might shift from being an integrated device manufacturer (IDM) — a company that designs, builds, and sells its own chips — to more of a foundry, which makes chips that others design.
Otellini says the company has no plans to offer its manufacturing capabilities to big competitors like Qualcomm and Apple. But, as many bloggers are quick to note, he is on his way out, and Intel has already made some movement in that direction by offering foundry services to a few select start-up companies.
When I spoke with VLSI's Hutcheson last year, he noted three things that make a foundry shift unlikely—at least in the short term. For one thing, the margins far less attractive. A foundry like TSMC simply doesn't make as much on each chip because it must split profits with its customers, the chip designers and sellers. For another, a foundry isn't just a chip factory; a good one offers a range of services to its customers, such as small-batch chip "tape-outs", to make sure the manufacturing process will work well and that yields will be high from the start.
Lastly, chipmaking is only getting more expensive and complicated. To get the most advanced chips into production, designers, design automation companies, and foundries now collaborate quite closely for years in advance of a chip going to market to make sure everything goes smoothly. The foundry space is "starting to look more and more IDM-like," Hutcheson says. And if that's the case, maybe the company that has been an IDM from the start will have the advantage.
(Photo: BenMargot/AP Photo)