Intel Inside? Not Likely.
By SoupIsGood Food, (soup@macedition.com), August 6, 2002
Feedback Farm
Have something to say about this article? Let us know below and your post might be the Post of the Month! Please read our Official Rules and Sponsor List.
Forums
Want to dig even deeper? Post to the new MacEdition Forums (beta)!
So, some yahoo with an investment banking firm wrote himself a manifesto. Fifty pages, it was, and it showed uncommonly poor sense.
First clue Andrew Neff is an idiot: He thinks Dell is big enough to rival its suppliers, Intel and Microsoft. Here’s a hint: To bolster its PC business, IBM made its own x86 chips for a time and rolled its own OS that was light years beyond Windows. In each instance, IBM got creamed by its suppliers…the Wintel duopoly. Dell doesn’t have a multibillion-dollar Unix and mainframe business to fall back on, either. While it may be turning its nose up at the Itanium (and who isn’t?), if Microsoft says jump, Dell’s already in the air. Dell is a huge juggernaut in the PC world, but such juggernauts rise and fall with depressing regularity. Anyone remember Zeos notebooks? Micron desktops? Of course not…and both companies were top-tier PC vendors. The only long-range winners in the Wintel game are Microsoft and Intel, and Intel’s on shaky ground.
Neff also believes that HP may set its sights on EMC next, and this has Dell in a tizzy. While EMC is great and good and has all kinds of crazy mindshare, it doesn’t change the fact that everyone and their pet monkey is putting out SANs and enterprise storage solutions. It’s not that hard to slap a FC-AL controller in a rackmount chassis, and if Dell wants to make its own printers and PDAs, high-end storage systems are probably on the menu, too, regardless of their association with EMC. Dell would rather keep all that profit margin to itself. So, Neff isn’t off to a rollicking start covering his home beat. But then he goes and steps into my territory, and now we’ve gotta rumble.
Second clue Andrew Neff is an idiot: He thinks Apple will be migrating to Intel chips in two to four years. Trouble is, nobody knows where Intel’s gonna be in two to four years. 64-bit chips? Perhaps, but which ones? Intel is pimping Itanium, but vendors aren’t biting, and Linus Torvalds himself wishes Intel would give up and go with the “Yamhill” x86 instruction extensions for 64-bit computing. Intel not only has no strategy to move to Yamhill, it outright denies it has any intention of getting one. Meanwhile, in the wonderful world of Itanium, I wonder if the oracular Mr. Neff would be so kind to explain why HP, Intel’s partner in the design of the 64-bit Itanium chip, is still developing new generations of 64-bit PA-RISC processors, if Intel is such a sure bet in the near future? In the meantime, Sun will be moving forward with clockless chips, and you better believe MIPS and the other small and hungry microprocessor design houses will be close on its heels. So, in two to four years, clock speed will be as meaningless on top-notch computing platforms as serial port speed is on an iMac.
And I wonder if Mr. Neff is aware that IBM has just opened up a new microprocessor fab – not exactly indicative of waning interest in slinging silicon on the open market. Waning interest in chips or not, IBM has some serious chops in the field. A dual-core G5 on a 0.10-micron process? How can that not be a winner? Also remember, the goal is “fast enough.” Macs are fast enough to run content-creation, productivity and personal software so the end user won’t be able to detect much of a difference between a 2.5GHz PC and a 1GHz Mac. Intel and AMD get smaller and smaller returns on their stratospheric clock speeds, while PowerPC “cheats,” using AltiVec and a damn competent RISC architecture, so it squeezes out more oomph per clock cycle…x86 systems are faster, but not that much faster, and not for all applications. This should not be news to anyone by now.
Don’t get me wrong – IBM and Motorola have let Apple down by not aggressively pursuing technical advantages for the PowerPC. Hell, they let themselves down, allowing ARM and MIPS to eat their lunch in the embedded space. But the PowerPC has a hell of a lot of kick in it…remember, IBM’s Unix and AS/400 servers all run on the same instruction set as the PowerPC. IBM still sells servers based on the 604e, and Unix software designed for it will run just swell on POWER4 chips, which are currently the all-time speed champs of the universe. Groupe Bull in Europe uses these chips in its own servers, so IBM is willing to share. It’s more likely Apple would lean on Big Blue to cut a price break on POWER chips than it would be to jump ship for a shaky Intel future. IBM’s RS/6000 was the first RISC computer on the market, and its descendants are still going strong today, forming the backbone of IBM’s server business. That’s a long-term commitment. Meanwhile, Motorola is selling G4s hand over fist to companies like Cisco, who need a lot of horsepower in a small, efficient package. While the chips that Apple is after might be slow to improve, improve they will, because the entire POWER/PowerPC line is so valuable to each company’s core business beyond Apple. Apple, meantime, is willing to absorb some of the R&D cost and effort. It’s a pretty good match, and the company won’t get that kind of cooperation from Intel.
With Intel Inside, the Macintosh would become just an expensive PC that’s crippled so it won’t run Windows. An evil fate for such a wonderful computer platform. What’s worse, that hare-brained scheme has no role model for success, and more than a few templates for failure. IBM, Sun and Pyramid (now marketed as RM/Reliant, part of the strange brew of Fujitsu-Siemens’ enterprise computing product line) – they all tried this same crazy Intel-based “proprietary platform,” and they all fell on their faces, while their RISC/Unix offerings are still going strong.
But, hey, Neff’s a big-shot tech analyst working for a huge brokerage, so he must know what he’s talking about, right? Those high-tech flying analysts sure called that New Economy thing…
