A couple years ago graphic chip maker 3dfx was so far ahead of its competitors that people often used 3dfx and 3d accelerator cards as synonymous. Its competitors tended to put out propriety chips that didn’t work well (if they worked at all) and had numerous problems. 3dfx was on top of the world and looked, to me at least, untouchable.
This week, of course, 3dfx went out of business with rival chip maker and reining 3d king Nvidia acquiring the only thing of value 3dfx had left — its intellectual property in the form of patents and some physical assets. What happened?
3dfx got greedy. Rather than simply turn out butt kicking graphics chips, 3dfx listened to the suits who were whispering in their ears that they could make even more money if they cut out the middlemen who put their chips on cards with different configurations and went into the business of exclusively producing graphic cards with their chips. So where once you could buy a 3d card with a 3dfx chip from any number of companies, now the only place to buy such cards was from 3dfx. Unfortunately for the company, it sucked at getting graphics cards out almost as much as it excelled at creating state-of-the-art graphics.
Nvidia positioned itself simply as a graphics chip company and won. Why deal with all the headaches of different card configurations and technical support for every iteration — sell the chips to third parties and let them deal with that hassle (and conversely, let competition among card makers give consumers a much wider range of options than Nvidia could possibly support on its own).
In addition, Nvidia embraced Microsoft’s Direct 3D while 3dfx’s Voodoo cards supported its own Glide system exclusively. Early versions of Direct 3D were horrendous (and some argue the API has improved little since then), but hardware companies ignore Microsoft at their own peril
Of course hindsight is always 20/20 and it’s always that easy to tell the difference between a boneheaded move and passing up a golden business opportunity, but everything 3dfx was doing was so far removed from its bread and butter graphic chip business that it would really have been a shock if the company had been able to pull it off.
There’s also a lesson in there about just how quickly a company can go from market leader to dead meat. A lot of people thought 3dfx was making the wrong decisions, but nobody to my knowledge suspected they’d be out of business in only two years, much less from Nvidia which prior to its TNT chip produced cards that were generally considered inferior at best.