David -
I'll try to address your concerns one at a time. I'm an employee, but I'm going to give you my personal, non-employee, non-company sanctioned point of view. I hope you'll treat it as such.
1. Graphics is still a strong point for SGI right now, but the question is, will people need SGI's for graphics applications in the future as the Intel architecture gets more and more capable? Doubtful.
It wasn't graphics per se that made SGI what it is today, it was primarily the manufacturing sector, which realized what gains in productivity were available through digital prototyping and testing. While it is true that some of the lower end tasks have become accessible to machines made by other companies, most of this market is still far out of reach for today's mass market machines. In addition, as the Intel architecture gets more and more capable, SGI will be right there with it, stretching the capabilities of that class of machine to the limit.
2. Sure, there will always be high-end users, but that market is growing slowly compared to others - Cray was no cash cow, and never will be.
It is true that the HPC market is growing slowly, but it's over $1B/year, and it's still growing. SGI owns over 50% (not sure of the exact %) of that market, and growing. Cray legacy machines are just what you claim they're not - a cash cow. All the R&D on these machines is done, all SGI has to do is keep making them and selling them, for whatever margin they will support.
3. This NT computer that you guys are hyping... how precisely is that going to help? Does anyone really think SGI will make money on a machine like that? They are just entering the market way too late. Are they really going to have comparable tech support to competitors? Be able to mass produce it? Are they used to marketing machines like that? It's a machine that will be a complete change of mindset from anything else SGI makes. No way this thing will be successful. How many companies have pulled off something like that, in any business? Now how many companies have royally screwed up doing the same thing? That would be like my specialty chemical company buying a pharmaceutical - it just won't work at all. I'm thinking the NT machine is a gamey proposition.
The NT machine is a way to get at the commodity market. Not only does this provide an opportunity to get SGI 'in the face' of the average Joe, but it's a great chance to make great graphics available in large quantities. Look at the paradigm shift in console gaming created by the N64. 3D in every kid's room. Neat, but it also taught SGI a lot about how to do graphics for cheap. NT creates the same kinds of opportunities. As for our tech support, I can't say. The OS support will obviously come from MSFT, with the exception of the SGI H/W specific pieces. I would point out, though, that our tech. support is quite superior to that of the traditional wintel maker. Ever tried to get low-level specs on a video card that ships with a CPQ? What a pain. IMO, one of the reasons this machine is being brought to market later, rather than earlier, is to make sure that the delivery, support, and reseller channels are in place before it hits the street. IMO, that's the way to do it.
4. MIPS - should be spun off, that might boost the stock up to my exit point! Whoo Hoo!
Pardon my saying so, but I think you're up in the night. Not only does MIPS make a big difference to SGI today, in terms of both mindshare and money, but the future potential is amazing. Think of what the 68k did for MOT. Now, think of a better designed, cheaper, faster, and less power consumptive chip. Today it goes in cameras, laptops, video cards, palmtops, settops, etc. Tomorrow, who knows? With the right moves, MIPS alone could propel SGI into the living rooms, cars, kitchens, bathrooms, and boardrooms of most of the people in this country.
5. Challenging SUN is absolutely hopeless. SGI clearly doesn't have the marketing oomph to do it, and they've already missed their window. That would have been the last 2 quarters of f&*k ups. SUN is one of the gorilla's, and SGI is just a chimp picking up the scraps.
This one is an uphill battle. I disagree that we've missed our window. I do think that we need to act soon, though.
6. What can Belluzo possibly do? Can SGI really change from a high margin model to a low margin model, and penetrate a market already full of well-branded competitors? Uhhmm... NO.
He can provide direction and marketing strength. He's also a new face that will allow what Vincent's been drooling for all along. Amazing graphics, available at affordable prices, for everyone. He has the opportunity to turn SGI away from a niche 'high end' company, into a company that can provide the same kinds of products to Joe User. NT makes this possible. It really doesn't matter what the brand on the box is, because Joe User views them all as being the same (Intel inside and all that). What matters is price and performance. Our performance has always been the best, what we really need is to become competitive on price. Corporate mindset has been a big hindrance in this area, and Mr. Belluzo has a chance to change that, if he wants to.
7. I must say, I would like to see an ad campaign, it will help the stock out, but for an irrational reason. I mean, it makes sense for Intel to do ad campaigns, since their market is the entire American population. But SGI's market segment? It seems awfully small to use a mass marketing tool. Even for the NT machines. I haven't seen too many Intergraph commercials. The only reason Sun and Oracle are advertising is because they can't stand to let Intel and Microsoft be the only companies to advertise to the masses.
And neither can we. Unfortunately, it's hard to let the general public know that a R12k @ X Mhz is faster than a Pentium<whatever> @ X*2 Mhz because of architectural differences. It's hard to counter Merced FUD, even though
o It's not a real chip.
o It's a re-implementation of several already-failed ideas.
o It's too complex to be widely workable.
o It's too hot to ever be used in a desktop, or even a deskside.
Joe Public doesn't care. He just cares that INTC says they'll ship a 1Ghz Merced CPU in 2000. That said, though, have you noticed that we appear to be doing quite a few more press releases as of late? |
