Paul I think have it exactly backwards. SGI has never, really, before been in the low-end. The Indy had price point at the $8,000 mark, but no usable Indy could be configured for less than $16-24,000 (certainly true a year ago). Usable O2's can be had for $7,500 which while more than your basic GW2K, is quite competitive with comparably performing Wintels. SGI **shipped** 10,000 O2's in the first quarter of availability, more than any 1st quarter product ever -- and built backlog for the product.
SGI's problems seem to be more in meeting and forecasting the demand, rather than simply creating one. I agree SGI needs to make its numbers to grow the stock price and give us investors a competitive ROI. The challenges appear to manifold (and the merger charges never-ending), but as for the product line, they can't seem to ship enough of them.
Hmmm, here's a Catch 22... if they have a backlog, they must be pricing their systems too low, as demand is exceeding supply, thus they are losing potential short term profits, miss their numbers, and we all complain, and the stock does flip-flops. If they raise their prices, the Wintel Chicken Littles of the world accuse SGI of being "out of touch"(BTW, the memory and peripheral pricing IS out of touch IMHO, but thats what Kingston and DataRam are for, right?) and cry out, "the windows are falling, the windows are falling."
So Paul, which is it going to be.
(a) SGI prices are too low. See their backlog? Look they missed their numbers again! (Then of course there is the cost of lost Goodwill for gouging customers...)
-- or --
(b) SGI prices are too high. Wintel is taking over! (Then of course you have to explain that pesky backlog...)
-- or --
???
I agree that SGI needs to "perform" (defined by market expectations), but the issues appears, at least to me, to be execution, not technology. (Soberly remembering of course, that the best technology doesn't always win, but often the best marketed does... but then you're back to the same contradiction. They can't build the systems as quickly as they are ordered, and you want to advertise more?!?) Long term, clearly SGI does need to build mindshare and volume to continue to push it's entry point down (note again that it is SGI that has move into the low-end, not vice versa).
On my wish list, and I think on that of many potential SGI developers/users, would be a GW2K priced entry IRIX system, just an R5K, 3-4 of PCI slots and drivers for: one or two hot > $500 PCI gfx cards (Stealth and Millenium for example), soundblaster audio, one or two PCI ethernet cards. Barebones + IRIX for 2-3K (at least for students and developers). The used Indigo 4k's and Indys are filling that need now, but the O2's are actually too good (and thus should hold their value used too well) to do that for quite awhile.
While the company is riding high in terms of current product demand, I don't see that happening (mostly from the opportunity cost of redirecting scarce "top" talent that can contribute to the bottom line better elsewhere), but it is always an option -- meaning that SGI could push even lower into the low-end, if they wanted or needed to.
I'm still long, but offering no investment advice, and speaking personal opinion only.
PS. An comp.sys.sgi.announce post said to watch for a "major" announcement next week -- oooo, the suspense!!! |
