It's Intel vs. Transmeta as Low-Power Chips Hit the Market
By Alex Lightman
internetworldnews.com
Like many observers, I was stunned by the relative absence of personal computers at PC Expo, held in late June in the sweltering heat of New York's Javits Center.
The white-hot heart of PC Expo was the center that held the dueling booths - set within rock-throwing distance of each other - of Palm, Intel, Handspring, and Transmeta. On one level, it appeared that challengers were trumping their spectacularly successful rivals: Handspring and Transmeta had the highest excitement levels of the convention and had plenty to show off, with Internet-enabled mobile devices being the clear imperative.
Transmeta made big points by offering cigar-shaped fans that offered the only relief from temperatures seemingly above 90 degrees. In case booth visitors missed the point of the whirling fans, there were thick TV totem poles showing infrared videos of the white-hot Intel chips vs. the cool blue-green Transmeta Crusoe chips.
Transmeta's offerings are built around two chips, the 333- to 400-MHz TM3200 and the 500- to 700-MHz TM540o, with only the slower chips currently available. The high-speed TM5400 will reportedly cost $175 in quantity - far too expensive to be included in mass-market handhelds.
Intel's booth had much less traffic and seemed stiff and uncreative in contrast to the tropical-themed Transmeta booth. All in all, it looked like the hype was real: Transmeta may well beat up on Intel in the market for mobile Internet devices, the most exciting hardware market on the planet.
The reality, however, is that Transmeta is still a very small company, and it can barely keep up with inquiries, let alone chip orders. As its Web site states: "The Crusoe processor design package is currently available only for Transmeta's early product design partners. Due to overwhelming demand for Crusoe processor design information, we will be making the Crusoe Design Package more widely available at a later date." If Transmeta's CEO, one of the busiest men in Silicon Valley these days, has to allocate precious design packages personally, Intel is much less vulnerable than is currently speculated. After all, Intel does more than $30 billion a year, posts annual profits greater than $7 billion, and boasts a market cap that recently topped $400 billion. Intel has occasional problems keeping up with supply, so imagine fabless Transmeta's challenge.
Intel knows, from nearly three decades of hard experience, how difficult it is to scale up to meet the demands of developers. If Transmeta has a temporary advantage for wireless devices in terms of market perception - based on low power consumption and unique abilities to deal with long instruction words - it's possible to fail to execute on all the other things that make a chip company successful, in a market where success against Intel means being among a scant handful of survivors.
In part as an attempt to cool down, and in part because of Transmeta's employment of Linus Torvalds, I thought about penguins, the symbol of Linus's Linux. On the icy rim of Antarctica, killer whales eat Emperor penguins, so the penguins don't jump in the water if they see a fin. They crowd around the edge of the ice, afraid that there might be a killer whale underneath. As Charles Kingsley Levy puts it in his book "Evolutionary Wars," "Penguins also play a cruel survival game in which one unwitting player is bumped off the ice floe to literally test the waters for the presence of predators." If it's eaten, they all wait, and, from time to time, they'll bump others in. "If the victim swims to the surface, it means that all is clear and the rest of the penguins dive in."
It occurs to me that Transmeta could serve the purpose of a test penguin in a competitive marketplace filled with killer whales. For example, before Intel goes out and spends $7 billion to make a new plant, it can do massive, free market research based on the demand that Transmeta generates. Thus the PC Expo dance could be seen as Intel bumping Transmeta into the water. If demand is huge, expect Intel to jump in with both feet in a bet-the-company focus on low-power chips that leave its native line far behind, perhaps based on the StrongARM technology, of which Intel is a co-owner.
The legacy of PC Expo will be Intel's eventual turning away from the PC and toward wireless devices, thanks to the brilliant pioneer-engineers at Transmeta.
--------------------------------------------------------------------------------
Alex Lightman is co-founder and CEO of MIT Media Lab spinoff Charmed.com, formerly InfoCharms, which is pioneering small platforms for Internet services. Reach him at alex@infocharms.com.
------------------------------------------------------------------------------
Copyright © 2000 by Internet World Media, A Penton Media, Inc. Company. |
