Intel told stockholders Monday that CEO Paul Otellini would retire in May after eight years at the helm of the world’s largest chipmaker. So far, no successor has been named. His decision to retire comes as Intel’s x86 architecture is facing a an existential threat brought about by a need for more efficient computing for both mobile devices and inside the data center.
From the release sharing the news:
“I’ve been privileged to lead one of the world’s greatest companies,” Otellini said. “After almost four decades with the company and eight years as CEO, it’s time to move on and transfer Intel’s helm to a new generation of leadership. I look forward to working with Andy, the board and the management team during the six-month transition period, and to being available as an advisor to management after retiring as CEO.”
It’s worth noting that Otellini is retiring at age 62 ahead of Intel’s general retirement age of 65. Otellini has had several roles inside Intel, and was only the fifth CEO at Intel since the chip firm was founded in 1968. While at Intel Otellini helped convince Apple to replace the Power PC chips inside its MacBooks with Intel processors. He also presided over Intel’s $1.25 billion settlement with AMD over allegations that Intel was abusing its power in the marketplace.
However, as the world increasingly went mobile, Intel has seen some of its power recede. It has adopted new form factors — remember the ultrabook? — but not new processor architectures. Instead, it tried to force the x86 architecture into a lower-power chip, which gave us the Atom chips. It also bought the wireless business of Infineon with the hopes of creating an integrated application processor and radio chips similar to what Qualcomm offers in the mobile space.
Currently the Motorola Razr i has an Intel-based application processor, the first mass market smartphone to have such silicon. But its success in this area is not assured.
The changing chip ecosystem
Earlier this month Qualcomm’s market cap surpassed Intel’s in what many people saw as an example of the mobile dominance over PC. But Intel is facing threats in the data center and high performance computing worlds as well. As far back as 2008 Intel tried to develop an x86-based graphics processor aimed at high performance computing, but its Larrabee efforts failed. It went back to the drawing board and is seeing some success with its new Xeon: Phi architecture, which is in a few of the Top 500 supercomputers and has the potential to keep Intel competitive in the market for massively parallel computing.
But at the other end of the data center spectrum Intel may have already lost. Efforts to use so-called “wimpy cores” — which offer less performance but sip very little energy — inside servers has opened up the data center to ARM-based chips. Even AMD, the only other holder to the x86 license, has taken an ARM license to make servers using ARM-based chips. That’s not to say Intel’s market share in the enterprise computing market will go away, just that there’s a much wider array of processors out there that infrastructure architects are eyeing for their computing jobs.
And that will change the economics of the chip business for Intel. Where it was once the keeper of the only game in town for mainstream corporate and consumer compute (yes, Sun and IBM have alternative architectures, but x86 was the mainstream) there are now multiple options for server makers and their end-customers. Many of those end-customers are even bypassing the server makers and have crazy ideas about renting CPUs and changing up board designs, that Intel will have to listen to.
What is core to the business?
So as Otellini leaves, his successor will have to adapt not only to more competition but also a different style of competition. ARM licenses its IP and thus, many different companies are innovating on that architecture competing on both price and features. Intel has some of the smartest software and hardware designers in the world but they will be competing more closely against outside firms on features, and Intel’s sales staff will have far more competition on price. They will also have different customers — both server makers and end users.
Aside from the competition angle, there’s Intel’s business model at the moment. Intel makes its own chips, conducts billions of dollars in R&D to keep Moore’s Law (named after Intel’s founder Gordon Moore) going and has to sell enough products at a high enough price to support its multi-billion manufacturing operation. Unlike Samsung, it doesn’t make chips for other companies, but that may be changing. In February Intel began a pilot effort to open up its manufacturing facilities for other types of chips
The opportunities presented by the changes in the chip world are incredible, and Intel has resources that few other competitors possess, which means that the right leader could take this role and use it to remake a chip powerhouse for the next decade. Our need for computing is only expanding, it’s just that the nature of our computing has become both more varied and the competition more fierce. The next big opportunities in chips aren’t in general purpose anything, so where will that leave Intel and x86?