← Back to context

Comment by cbm-vic-20

10 hours ago

I had a fair amount of shares of TDFX back in the day, rode it all the way to their bankruptcy. NVDA bought the scraps and left retail shareholders like me with nothing. In a different reality, if 3dfx could have survived to the AI-era...

Really nice background in the recent book The Nvidia Way. The gist of it was Jensen recognized that having the best chip once wasn’t enough.

The result: Nvidia organized around delivering something new every six months. This meant a new architecture followed by two improved derivatives over an approx 18 month period. That turned 3dfx’s development schedule into a trap.

Conclusion: disciplined execution above all else.

Its interesting how even I didn't realize how competitive he was...until after the market realized recently. Wish these books would come out a year earlier than they did. :/

I guess i'm not cut out for stock trading. I'm always arriving as the train is departing both in real life and on the market.

  • > Nvidia organized around delivering something new every six months

    That explains the SUPERs that were released seemingly very quickly after the original models. Looks like the 5 series will miss out on that due to the ram shortages and prioritising their HBM server models.

I still remember playing Quake 2 to death in software then finally getting a 3dfx card (I want to say it was a Banshee?), loading up the game and my mind being blown by how much more immersive it was.

Great times, the 3dfx name will always raise a smile for me.

  • Core memory unlocked. I remember being in my dark bedroom with he glow of the CRT playing Quake 2 grapple CTF for hours and hours. Eventually I got a Voodoo 3 in all it's 16-bit color glory.

    It's wild how hazy those memories are becoming. Getting old is weird.

    • Same. I spent an insane amount of my time practicing grappling in Quake2 Weapons Factory (and eventually WFA). I remember getting the Voodoo3 3000 seeing how amazing everything looked, and then immediately turning it all back off to get the highest FPS possible. Considering this was all over dial-up it's funny to think back on it now.

    • Oh my god....q2ctf was the absolute best. I still fire it up on LAN parties here from time to time.

I wanted a 3DFX card so bad when I was 10yo.

And then I wanted a GeForce so bad. But I only got £10 a week from my paper round.

If I’d invested just one week of my paper round in Nvidia shares though, it would be worth over £10 million today.

  • When I was a kid I told my parents I wanted to invest some of my money in NVDA after we had a school assignment to study a company. They laughed and gave me a speech about stock picking. What they said was actually good advice, but looking back it would have been a home run.

    At the time stock trades were still expensive. I remember one of the obstacles was thinking that of the small amount of birthday money I had saved up, $20 was going to go into the purchase and another $20 would have to come out of the sale. Then they explained taxes to me.

    None of it would have mattered if I had been able to buy and hold, but at the time I remember being sad that even doubling my money would result in $40 of fees plus taxes, eating most of the gains.

  • More like 2 years of your £10 a week paper round for that £10m return.

    • Their stock was $12 on debut but they’ve had many many stock splits since then. So the equivalent price of a share is less than a dollar. And it went to the equivalent of 7 cents when the bubble burst.

Me too. I lost all my money at the time betting on TDFX. I was salty as fuck against Nvidia for decades until I had no choice but to buy their GPU 10-ish years ago. Now I’m a stockholder. I only wish I chose properly 25 years ago…

> In a different reality, if 3dfx could have survived to the AI-era...

...they would still have fumbled it.

3dfx failed for a single reason: Failure to design for the future.

They had a hyperfocus on brute force rasterization and fill rate that became detrimental. They dropped the ball on 32-bit rendering because they thought having a higher fill rate was more important. They dropped the ball on implementing hardware transform because they considered that the CPU's job.

And then they blew up their own business model buying STB Systems and manufacture cards themselves, rather than selling chips to other manufacturers. That ended up being a gift to NVIDIA who adopted the opposite strategy of providing chips, reference designs, drivers, etc. and let a large number of board manufacturers compete to sell them.

And then there's the fact that NVIDIA was allowing themselves to create new architectures, while 3dfx was basically just saying "Let's make the same GPU, but with more texture units". By the time the Voodoo5 came out, NVIDIA was already on the GeForce 2.

3dfx lost because it kept optimizing for yesterday. NVIDIA understood where the industry was going and 3dfx didn't, and there's little reason to think AI would have gone any differently. If 3dfx had somehow survived another 20 years, I'd expect them to arrive at the AI boom with the world's fastest rasterizer.

I remember watching those promos of need for speed 2 SE running on a 3DFX and feeling like a peasant for not having one.