On December 1, Intel’s board of directors named longtime Intel insider Justin Rattner as the company’s new chief technology officer – and only the second CTO in Intel’s history.
Even before his elevation, Rattner, 57, made frequent public appearances on Intel’s behalf (including a presentation at Technology Review’s Emerging Technology Conference in September 2005). Indeed, he’s widely viewed as the company’s most charismatic geek, with an infectious enthusiasm for Intel technology that earns him comparisons to Apple’s Steve Jobs.
Rattner will need every ounce of that enthusiasm to carry Intel through the next few years, as it attempts to replicate the success of its Centrino wireless initiative and transform itself from a maker of computer components into a “platforms” company that markets entire packages of computing and communications technologies for homes and businesses (see “Intel’s Centrino Solution,” February 2005). And Rattner’s charisma will certainly be tested as Intel herds together companies around emerging computing and communications standards such as WiMax, expands into new markets in the developing world, and coordinates its R&D efforts with business units, as the company plans for its next generation of microprocessors.
Rattner, who joined Intel in 1973, is one of just 10 Senior Fellows at the company, an honor given in 2002 in recognition of his work on high-performance computing, especially distributed and cluster computing. He has led Intel’s research operations since a major reorganization of the company in January 2005. Technology Review’s Web editor, Wade Roush, spoke with Rattner on December 1.
Technology Review: Before your appointment as chief technology officer, you were already the director of Intel’s corporate technology group, where you led Intel Research as well as the company’s microprocessor and communications systems labs. Will you keep that position? How will your job change now that you’re CTO?
Justin Rattner: I’m going to continue to lead the corporate technology group. I’ve been doing a lot of the “external interface” things for Intel, working with CTOs of all the leading companies and being part of the public presence for Intel technology, and that’s certain to expand. But I think more of the change will be less visible.
I’ll have to do more to balance the longer-term research interests of the corporate technology group with the broader business interests of the company. I’ll be more concerned about the technology decisions taking place over the near term than I would normally be in the corporate technology group. So we’re talking about me being greater public presence and taking a broader look, and being perhaps the conscience of the company in terms of technology choices.
TR: You are only the second chief technology officer in Intel’s history; Pat Gelsinger occupied the position from 2001 to 2004, then it was vacant for almost a year. Why does Intel need a CTO?
JR: That’s probably a better question for Craig Barrett [chairman of Intel’s board] and Paul Otellini [president and CEO]. But I think it really reflects the broadening of Intel’s technology agenda. In the earlier phases of the company, where we had a narrower focus on memory chips and then microprocessors, it wasn’t really necessary to have one person trying to represent and articulate the broad [technology] interests. Quite frankly, you had some of the legends running the company – Bob Noyce and Gordon Moore. They were very effective as CTOs, even though they didn’t carry the title.
I think when we were moving into the new century, the view was that the technology issues have broadened and become more complex. There are now a lot more policy-related questions and standards issues covering both computing and communications. There was a real need – almost an expectation within the industry – that there would be an individual at Intel that you could go to for that whole range of specific technology questions. There was a recognition that it was beyond the ability of the CEO or any one business group manager to cope with.
TR: What’s a concrete example of those policy and standards issues?
JR: WiMax [a standard for metropolitan-scale wireless broadband communications] is one example. The broader picture is just spectrum policy in general, and making spectrum allocation consistent on a global basis. Advocating for the different technologies at the industry level is clearly a big part of it, whether you’re talking about WiMax, wireless USB, or any of the ultrawideband technologies, or developments like multicore processing. There are any number of technology inflections coming, and Intel can’t single-handedly drive all of these major transitions. You have to enlist the industry, and I think that’s clearly on of the things the CTO, in particular, can be a focal point for.
TR: How will you tackle the areas where Intel’s been facing difficulty lately, such as delays with the introduction of your dual-core Itanium processor for servers, or the fact that desktop PCs running on AMD’s Athlon chips have been outselling PCs with the Intel Pentium 4?
JR: I think it’s really an issue of being able to put alternative technologies on the table, to give the company the options it needs to make the right decisions. It’s fine to say that we’re running out of gas on the current approaches [to chip design] and we need to make a change. It’s another thing entirely to actually look at moving from what we’re doing today to the future approaches and alternatives. Part of my responsibility is to make sure we’re putting those technology options on the table.
The other thing is to make sure that there is a real sense of collaboration with the business groups, in terms of bringing the new technologies into the product line. You can’t just throw these things over the wall and say, ‘Here’s a proof of concept, you figure out all the details and call us back.’ Increasingly, corporate technology [researchers] at Intel and our digital enterprise groups have joint laboratory activities, where people from the development side and the research side are working together to make sure that we make these transitions successfully – and not just successfully, but on time.
TR: Intel’s view has been expanding outward, toward developing markets. What are your most important international initiatives right now?
JR: That’s what I spoke to at the Emerging Technology Conference. We used to think about products primarily for U.S. markets and American consumers, [but] that’s a very small part of the global market. The other assumption people had was, ‘Well, if it’s good for the U.S. customer, be it a business customer or a consumer, then it will be just fine in Asia or Latin America or the Middle East.’ It’s become quite clear that that doesn’t work, and isn’t going to work going forward.
So we’ve basically expanded our skill set over the last four or five years to include the kind of expertise that it takes to understand how culture, living standards, and even climate affect the way systems are designed. That’s what’s driving the regionally optimized platforms we’ve introduced. That’s only going to grow – it’s not a one-shot deal. There’s nothing like getting products out there and seeing how people react to them.
TR: What’s your favorite example of one of these regional platforms?
JR: Well, the China Learning PC. [The China Learning PC is a prototype computer designed to address concerns expressed by Chinese parents that kids will use home PCs to waste time. It comes with a key parents can use to lock the machine into regular mode or “education” mode, restricting access to non-education-related software and websites - Ed.] And I think our ICafe products [for managing large numbers of PCs in Internet cafes, also being tested in China- Ed.] are hugely popular. Internet cafes are not a phenomenon here in the U.S. to the extent they are internationally. That’s clearly a product that appeals to non-U.S customers, and it has satisfied the needs of both users and operators, the people who run these Internet café businesses. It’s providing a perfect example of how you take a concept like that, bring it into a specific market, and use the learning for subsequent generations of design.
TR: Now that you’re CTO, what do you hope to accomplish first?
JR: Good question. We should really talk again in 90 days. Those first 90 days are the critical ones for anyone in a new job.