Recently received Email you all might find amusing:
Subject: silicon investor, graphics chips
From: jnjf@ix6.ix.netcom.com
To: ajberger@aol.com
Date: 97-06-26 16:06:26 EDT
Greetings!
I thought I would email you directly because I am still looking at joining the silicon investor site and haven't yet sent in my check. [you have to pay now?] Since you can't post until you pay, I thought I would give you a ring since you are moderating [now I'm a moderator? what does that pay?] the 3dfx channel.
[I cut out a lot of backround noise, here is the rest of it intact.]
3DFX: Here is a company that I think has learned from past mistakes
made in the industry. Current Voodoo solution is TWICE as fast as
the competition. Check out the benchmarks. They courted the
commercial video game guys (midway-williams, etc.), the home console
guys and the PC guys. Asked them what THEY wanted. They
designed the chip and the boards and gave them out to all the
developers from day 1. 350 developers I believe. At the E3 show,
they DOMINATED. Eyewitnesses say that 90% of the titles were
sporting 3dfx support. Rendition and NEC powerVR picking up the
rest. Well what do I think about 3dfx? They will hold their
own for a year. Longer, if they don't drop the ball. Their current
solution has enormous support and they are leading in performance by
a good margin. Rendition has announced their 2nd generation Verite'
chip. Twice as fast as current solution. That's nice, but, that
just equals current Voodoo solution. Probabaly are in second place
in vendor support. The way I see it, they need a quantum leap in
performance before people take notice. Okay, so you say that the
developers are not going to start individually supporting every
single chip that comes out? You are right. Very few developers took
the time to optimize their 2D games for every chip that came out.
Too hard and would take too much time to do. Instead, used a basic
layer called VESA to handle that. 3D will probabaly even out as far
as feature set goes. Developers will want standard lighting effects
and so on. Okay, then what do I buy? Gonna come down to brute
force power in a chip again. What differentiated the 2D guys. Not
features. Everybody had VESA video modes. Pure 2D power.
Currently, Matrox is champ. The basic layer for 3D is been fought
by OpenGL and Direct3D. Again, this will even out the feature sets.
This will probably take a year. In the meantime, 3dfx has the
advantage. Everybody will try to match performance and features with
them.
Okay, lets look at the other 3D solutions...
As I said, nobody is even close to 3dfx at the moment. Tritech
Electronics has a very good chip called the Pyramid 3D. Good
effects, and good performance plus they have the backing of their
parent company, Singapore Technologies. A Billion dollar parent I
might add. Good mousetrap, but very little support. We will watch
these guys... ATI Rage Pro? Still a joke..In the same camp as S3
Virge GX2. Mediocre features and mediocre performance. Aha! 3Dlabs
you say! Well the Glint boys are trying with their Permedia 2 2nd
generation chip. (Brian Hook at Id software has some intersting
comments regarding this and PowerVr chip.) This chip is still not
there. No support really and average performance.
Finally...
3dfx has their current Voodoo solution. Next, Voodoo Rush was
introduced to work together with a 2D chip. Same performance as
current Voodoo. Started shipping last month to OEM's. Voodoo 2
rumored to be 2nd generation chip. Talking about 4 times the
performance. TSMC, which is their Taiwan manufacturing partner, is
moving to a 0.35 from a 0.5 micron manufacturing process. Will
probabaly see shrinking die size, upped clock speed, and integration
to one chip with this Voodoo 2. Also working on Banshee, which is a
complete 2D/3D solution due to ship in Q1 98. They are covering all
the bases with this strategy. Very wise indeed...
Lets talk about Sega...
Sega is of course working with 3dfx for their upcoming "Black
Belt" console. It makes no sense for Sega to drop out from
the console market. They have the software arm and commercial
coin-op arm in addition to consoles. Why not sell a title 3 TIMES
instead of once. 3dfx solution was designed to allow easy
portability between platforms. Sell it on the console, to the PC
market and finally at the coin-ops. [This is an interesting arguement for SEGA to do another console, however, if NextGeneration has not picked up on this 6 months before Xmas, then I doubt its gonna happen]
Let me talk just a bit more about 2 cards versus one card
solutions...
I don't see what the concern is about. Currently, you can buy a
3dfx card (Orchid) for about $155. Ok, now I need a half decent 2D
card. (Don't fret about screamin' 2D performance, since 2D is dying
fast.) Get yourself a Diamond Stealth 64 Video 2201 2MB PCI DRAM
card. It has the S3 Trio v+(765) so you avoid the dumb memory
problem with the other S3 chips (864,868,964,968). What's that cost?
It runs about $70. Total= $225. Very close to sweet spot of $199
that most consumers can tolerate. Why get caught up in a ONE card
solution when Hercules and the other guys are working out the bugs.
Okay fine, probably $30-40 cheaper. That's nice. You can't afford a
dollar a day to pay for the difference in a months' time? Come on!
A can of pop a day costs more! I think my biggest complaint with
customers is that they can be penny wise and pound foolish.
Conclusions...
Overall, I think 3dfx is close to turning the corner to getting
in the black and think that they have a very good chance for
execellent appreciation in the next 12 months. I put my money where
my mouth is and bought the stock at the IPO yesterday.
If you want to post this in the channel, no problem. All that I ask
is that you post it all so that it stays coherent.
At least I hope its coherent... :)
[sorry to cut the first part out, but it was pretty lame]
Thanks for your time. |
