The Fallout Over 3Dfx
The deal to buy STB has its finer points. Here's what's behind the big deal.
I've been mulling over the proposed merger of STB and 3Dfx for a few hours now, and it's slowly dawning on my somewhat befuddled brain that there's a lot more complexity here than meets the eye. Most of the punditry so far has focused on either the obvious synergies or the obvious problems - which is the easy stuff.
Obvious synergies include giving 3Dfx control of its destiny - something it didn't really have so long as it had to peddle chips to every board maker in the known universe. Within limits, it will now be able to control price. Certainly prices will drop over time, but we may see 3Dfx attempt to control its chip prices the way Intel controls the pricing for its CPUs. STB also has inroads to a number of big OEMs, including Dell and Gateway. The merger makes 3Dfx look more like an ATI or Matrox in this respect.
The obvious problems include 3Dfx's relationship with other board companies. Even if 3Dfx hadn't said it will only offer Voodoo3 on its own boards (assuming the merger goes through), other companies will be far less interested in buying parts from a company they compete with. You can see this with ATI, who also markets graphics chips as well as full board solutions. But only a few Taiwanese companies offer boards with ATI parts.
But there's more here than meets the eye. One reason STB has good relationships with OEMs is choice. In the past, STB was able to go out and find the best solution to meet a particular customer need. A good example of this is the Nitro3D used in Gateway's Destination product. This card used the now-moribund Chromatics Mpact II accelerator to handle 2D, 3D, and DVD chores. Now STB can't meet particular customer needs if it lies outside the purview of 3Dfx products.
In addition, it could distract 3Dfx from its core business. Now 3Dfx will have to meet the requirements of all those big OEMs up front. And woe be to he who cannot run Winbench as fast as the competition. Passionate adherents to 3Dfx products will point out that its upcoming chips are plenty fast, but the road to graphics Nirvana is paved with the silicon of accelerators who formerly held the title of fastest kid on the block.
Some people are already sounding taps for nVidia, but I think that's premature. If anything, this could give a short-term boost to nVidia, as other board makers flock to the current fast kid. I suspect board makers will shy away from 3Dfx rapidly in the next few months. Why give your future competition any edge? And hot on the heels of nVidia is ATI's Rage 128, which now proves there's no real performance hit for true color, the Permedia 3, who may be a high flyer, and S3's upcoming Savage3D II.
The flip side of this coin is that the consolidation may continue. Already Rendition has been swallowed by Micron. If nVidia merges with a board company, then companies like S3 and 3Dlabs may find themselves lacking top-level vendors to move their products. 3Dlabs, of course, is already dipping its toes in the board business, having acquired high-end board maker Oxygen.
One thing is for sure. Just as everyone thinks he's beginning to understand the competitive landscape, some fault line shifts, and we're all left trying to figure out what happens next. God, I love this business.
Got a news tip or want to contact us directly? Email news@gamespot.com
Join the conversation