Apple CEO Steve Jobs sent a shockwave across the technology industry in June when he announced Apple would phase out PowerPC chips and put Intel processors inside Macs starting in 2006.
To some, the move seemed puzzling: Why would Jobs, the king of cool design, make a deal with half of the empire that conquered the world with beige boxes? Jobs had an answer at the ready during his Worldwide Developers Conference keynote - a switch to Intel chips means better Mac hardware down the line. And analysts agree that the move ensures Apple's ability to craft unique designs.
But one aspect of the "Why switch processor suppliers?" question hasn't been answered. Intel isn't the only X86 chipmaker in town. Why didn't Jobs, ever the maverick, opt for the scrappy challenger, Advanced Micro Devices, instead of the old-money establishment, Intel?
The reason, industry analysts say, is that Jobs has a clear goal in mind: innovative designs. And such designs require the lowest-voltage chips, which IBM and Freestyle were not going to make with the PowerPC chip core - and which AMD has not yet perfected
"This is a practical, pragmatic Steve Jobs decision," says Shane Rau, programme manager, PC semiconductors for IDC. Intel serves up the most complete line of low-power chips for mobile and small form factor computers, and a good-looking future roadmap for it. Also, Intel's mammoth production capacity erases any supply worries.
Intel's inside advantages
Mac users have come to understand that Apple had good reasons to wave PowerPC goodbye. The company knows trends when it sees them: mobile computing has become a way of life for many consumers. Yet PowerPC chips aren't travelling along this route. Apple needs faster chips, with more room to grow, and a chip partner with a clear roadmap for the future. Otherwise Wintel PCs could run too many miles ahead of Macs in the performance race.
Still, that doesn't explain how AMD lost out to Intel. AMD has made a name for itself with super-fast machines, especially popular with gamers and bargain hunters, who value the couple hundred dollars you can often save by buying AMD-based PCs instead of Intel-powered ones. Jobs may have liked AMD's hard-charging rep - but it's possible he saw some problems he couldn't ignore.
"One of the biggest considerations for Apple was getting a roadmap in all possible markets where they may play," says IDC's Rau, "and if you look at AMD's product line, there are some holes." Most notably, AMD hasn't invested in creating a line of low voltage and ultra-low voltage processors that competes with what Intel offers.
AMD would need to develop a chip core especially suited to low-power, as Intel did with the Pentium M, a costly undertaking. Plus, the overall sales opportunity for such chips isn't huge yet, says Nathan Brookwood, principal analyst for Insight 64. Because AMD's research and development budget pales next to Intel's, AMD has to pick its battles with Intel carefully - whereas Intel makes chips for almost every market niche. "Intel can afford to dedicate the resources," Brookwood says.
By choosing Intel, Apple gets access to the highly-anticipated chip code-named Yonah, a low-power chip with a dual core processor, which aims to band together the power of two regular chips. Aimed at notebooks, Yonah should arrive in PCs in the first quarter of 2006; in keeping with its tradition of remaining tight-lipped about future products, Apple has not commented on when Yonah might show up in its mobile line.
"Yonah could have been the tipping point for Apple," says Kevin Krewell, editor-in-chief of the Microprocessor Report. Yonah can power Apple notebooks that fly past today's models.
AMD does not have a direct Yonah competitor that would be available in the same timeframe that Intel is discussing. Is AMD working on a Yonah-like competitor? AMD won't discuss timeframe or specifics, but the company is currently developing a low-power, dual-core chip for thin and light notebooks, company spokesman Damon Muzny says.
Intel also employs a huge cadre of programmers, a resource that could be important to Apple as software gets rewritten for the x86 architecture, says Microprocessor Report's Krewell. AMD's programmer ranks don't compare in size.
Future AMD opportunity
Interestingly, performance really isn't the driving force behind Apple's Intel vs. AMD decision. While the chip rivals have battled on performance for years, the machines now go toe-to-toe on everyday productivity applications. For most consumers on the PC side, the buying decision is much more about the PC maker than the chip supplier. (That said, on some measures, AMD shines. Gamers, for example, who want the absolute fastest speed on traditional apps know that AMD's single-core Athlon 64 XP FX chips offer an edge over Intel's best right now.) As more multi-threaded apps designed to better take advantage of dual core CPUs arrive, Intel and AMD will keep battling.
Dual-core chips, which both AMD and Intel are emphasizing, marry two CPUs together for horsepower, but can share certain parts like caches and buses. Unfortunately, the dual-core chips are currently throwing a lot of heat, so both CPUs cannot operate at their maximum clock speeds.
Intel will tackle this problem in the second half of 2006, revising its product line with a new generation of lower-power dual-core chips code-named "Merom" for mobile, "Conroe" for desktops, and "Woodcrest" for servers. Intel will emphasize low power consumption and performance, but not megahertz, Brookwood says. (AMD has emphasized performance, not megahertz ratings, for years.)
"Intel seems to have kicked the megahertz habit," says Insight 64's Brookwood. "It's probably music to Steve Jobs' ears," he adds, noting how Jobs had to explain PowerPC chip performance on applications, not raw megahertz ratings.
Might Apple turn to AMD for future processor needs, post-transition to the x86 architecture? An AMD low-power chip line would be required for Apple to consider a switch, Brookwood says. But Intel will have a production capacity edge for at least a couple of years, an important factor, so a switch seems unlikely before then, Krewell says.