SAN FRANCISCO – It was a heady week hanging out the
Intel Developer Forum and covering the
iPhone 5 launch a block away. At the end of it all, I am both impressed and worried about the two big ecosystems of electronics that passed within spitting distance of each other.
Riding the big escalators at Moscone West I saw Dave Ditzel. He designed a couple generations of Sparc chips at Sun Microsystems back in the day and now is working for Intel on a CPU beyond anything the PC giant is talking about publicly.
Dave said he was impressed by Intel’s long, deep pipeline of microprocessor projects and its methodical process for executing on such massively complex programs. Me, too.
While we talked, the chief technologist of Invensas who also knew Dave chimed in, excitedly sharing a prototype in his pocket of the company’s latest 3-D stacking technology. For a moment I felt like I was standing at ground zero of the future of chip design.
I had that feeling again listening to Mark Bohr present Intel’s process technology road map. Bohr’s been around the chip fabrication business for 30 years and unarguably is one of maybe a dozen people now at its vanguard. He talked with authority not only about Intel’s next generation 14nm process but
the 10nm one beyond it. The whole electronics industry depends on pathfinders like him.
There’s an iceberg field ahead of the big Intel cruise ship. The lithography methods used to create chips seem to be running out of gas as we approach the atomic limits of scaling. Even Intel may not be able to stay much longer on its heady two-year cadence for new processes.
Then there’s the whole mobile thing. Yes, Intel’s Atom-based SoCs now power six run-of-the-mill smartphones and four compelling Windows 8 tablets. But Apple’s iPhone franchise (and the Android fleet led by Samsung) is steaming half an ocean ahead of it.
I did not get a feeling at IDF that Intel is in the vanguard of Android, and Microsoft is certainly not leading the way in mobile software. Intel has its work cut out for itself navigating it’s way to a strong post-PC ecosystem.
The good news is, it is still at the center of the PC and what people loosely call cloud computing. Jim Pappas, one of the guys who drove PCI and USB back in the day, is now enthusiastically telling me about how flash memory will change the programming model of the compute infrastructure, a model Intel aims to author—if all goes well with the post-NAND generation.
Meanwhile, Apple’s stock is rising on little more than the design of what amounts to a single, me-too phone. The iPhone 5 catches up with what its rivals have been delivering for a year—LTE, bigger displays, better media, smaller components like docking connectors.
It’s a great business model: Design one decent, but not bleeding-edge smartphone, tablet and notebook a year. Give consumers something that’s arguably cool and useful, and avoid the risks of bleeding edge technology.
Apple too could run aground, as Van Baker of Gartner pointed out to me in a chat at IDF. If it misses just once on one of its big annual product cycles, it could be hosed—losing a year of revenue, profits and panache as the company that sets mobile fashions. The risks are pretty big given the complexities in the software stack, the semiconductor and systems design chains and just plain human enterprise at this scale. Tim Cook, no doubt, buys Maalox by the case.
At the end of the day, it was humbling—and just a bit frightening--to see these giant cruise ships pass so closely as they sailed down San Francisco’s Moscone Street.
Related stories:Analyst says iPhone 5 processor is dual-core Cortex-A15
Intel describes 22-nm SoC process, not chipsIntel’s Haswell shines, but Atom remains cloudy