Interest in Intel's New Itanium Takes a Dell Hit
By Jim Seymour
Special to TheStreet.com
07/02/2002 01:54 PM EDT
If you're looking for some good news about Intel (INTC:Nasdaq - news - commentary - research - analysis) -- now stuck in the upper teens after its brief dance above $30 two months ago -- well, this isn't it.
Intel's new, second-generation Itanium chip, which the company expects to roll out next week, may prove less interesting to its OEM customers, at least at first, than Intel and its investors have hoped. Dell (DELL:Nasdaq - news - commentary - research - analysis), for example, has decided to pass for now on producing Itanium 2-based servers.
Even those other server makers who reportedly will sit at the table with Intel next week at the rollout -- IBM (IBM:NYSE - news - commentary - research - analysis), Hewlett-Packard (HPQ:NYSE - news - commentary - research - analysis) and Fujitsu -- are unlikely to rush machines to market or achieve much volume anytime soon.
Weighing the Significance
For Intel, Itanium 2 is important for at least four reasons:
It can clear up lingering unhappiness over the original Itanium chip released a little over a year ago, which didn't provide the expected performance jump and which, because Itanium chips require newly rewritten software, wasn't widely used.
Itanium 2 can be a bulwark against Advanced Micro Devices (AMD:NYSE - news - commentary - research - analysis), which has been touting its new Hammer CPUs as the better choice for next-generation high-power Intel-based servers.
Servers built around Itanium 2 CPUs -- often in clusters of up to 16 CPUs per machine -- are Intel's Great White Hope against Sun Micro (SUNW:Nasdaq - news - commentary - research - analysis) in the powerful-server market. Next week, look for Intel to provide comparative test data showing Itanium 2 servers match or outperform roughly comparable Sun servers for a lot less money.
Itanium 2 will be an expensive chip, and while Intel certainly isn't going to make any money on the first x-thousand Itanium 2 chips sold, it desperately needs to add more high-priced chips to its product mix to increase, even slightly, its average selling price across the line.
So why wouldn't Dell and, by extension, other server makers be falling all over themselves to rush out Itanium 2 servers?
Demand. Or rather, the lack of it -- at least at first.
Itanium's Future
Most observers see Itanium 2 machines as extremely low-demand boxes for some time to come. Their multiprocessor design -- single-processor Itanium 2 servers don't make much sense -- means interested customers are going to be taking a long, slow look first to make sure the bugs are worked out.
Someday, Itanium 2 servers may be popular, high-volume products for their makers. But I don't look for Intel to achieve any kind of interesting volume on Itanium 2s this year, and maybe not next year. The big server makers, including Dell, want big-volume products they can make cheaply and ship in large numbers. But there is likely to be very little in the way of economies of scale in building and selling Itanium 2 systems for quite some time.
If you want to make a technology statement -- "Hey, we're hip, we're hot, we're happening!" -- then sure, invest a few tens of millions in getting out an Itanium 2 line. Be prepared to work out the multiple-CPU software problems with Microsoft (MSFT:Nasdaq - news - commentary - research - analysis) and the clutch of Linux players who see Itanium 2 machines as the generation of servers that finally proves Linux's superiority.
But for most server companies, those self-indulgent days are over. And at Dell at least -- which would argue that it was never in that trendy crowd -- low-volume prospects mean a clear no-go on new products.
This is an awkward moment for Intel, but it can't back away from what has become a surprisingly strategic product, at a time when it needs some new strategic wins.
The Itanium may well pay off for Intel someday -- maybe in Itanium 3, which I believe will be the moniker eventually assigned to what is now known at Intel as "Madison," a sort of second-and-a-half-generation Itanium chip now in Intel's labs.
Does this provide some protection, for now, at least for Sun? I doubt it. Buyers interested in moving to the next generation in server power/performance will likely just put off that buying decision, rather than decide to give up and buy more Sun servers.
If this current corporate capex spending sag has taught IT managers anything, it's how nearly infinitely new-technology decisions can be delayed. And that's a dangerous new reality for technology companies across the board.
--------------------------------------------------------------------------------
Jim Seymour is president of Seymour Group, an information-strategies consulting firm working with corporate clients in the U.S., Europe and Asia, and a longtime columnist for PC Magazine. |
