Now lost in the annals of IT history, there was a time when 3dfx was synonymous with cutting edge 3D graphics on PCs. John Gunn found himself in the channel sales and marketing vice president position at 3dfx through STB Systems, a developer and manufacturer of video cards for OEMs such as Dell, Hewlett-Packard, and Gateway. According to Gunn, STB Systems had its own retail brand and developed products using graphics technology from 3dfx, NVIDIA, ATI, S3, Cirrus, NEC, and other manufacturers. For a graphics application-specific integrated circuit (ASIC) developer such as 3dfx, unless its technology was going on a PC motherboard, it went straight through STB Systems and Gunn.
You worked at 3dfx during a period which many consider the "golden era" of PC gaming. What was it like at the company during those heady days?
Former 3dfx channel sales and marketing vice president, John Gunn (JG): Overwhelmingly exciting. 3dfx's Glide technology brought arcade quality graphics to the PC platform and gamers freaked out over it. It created a new category of products for hardcore gamers. These weren't kids playing games on a console. These were men aged 18 to 40 with college degrees, and high average household incomes. Nothing else compared to it in quality or performance.
3dfx was instrumental in ushering a new age of video gaming through its graphics accelerator technologies such as Voodoo. What do you feel enabled the vendor to become the market leader and technological revolutionary in such a short space of time?
JG: Huge and immediately noticeable difference in performance. When you loaded top PC games such as Quake, during set-up you could select if you had Voodoo graphics. With Glade, the user experience was on an entirely different level, whether it was resolution, frames per second, visual effects. You either had Voodoo or you weren't a hardcore gamer.
How closely did 3dfx have contact with video game developers at the time?
JG: 3dfx was closer to game developers than anyone. Makers of expensive games knew that most of their target customers had Voodoo and they created a very efficient ecosystem for gaming on the PC platform. Unfortunately for 3dfx, Microsoft put more resources behind Direct3D and eventually caught up.
One of the more unique events in the company's history was its initial partnering with Sega on its Dreamcast console. Is it true that Sega had devised another Dreamcast candidate console, titled "Blackbelt", designed in the United States using 3dfx technology?
JG: I worked on the retail side of the business and was not a part of this. My perspective is that Sega initially thought that it wanted the best technology, then changed strategy and made a decision based on cost to hit a lower target shelf price for the console and culture.
Out of curiosity, do you have any idea why Sega terminated the development contract with 3dfx?
JG: Unfortunately, I don't know.
When you saw the PowerVR Dreamcast launch in 1999, what was your first impression of it?
JG: The industry gave a collective yawn. It was a decent solution, but not enough to get people to switch platforms. Other console developers learned from Sega's missteps and you have seen technology leap with the Xbox or the user-experience evolve with the Wii to drive every console launch since this time.
Could you speculate on what a 3dfx-based Dreamcast could have done differently than the PowerVR one that was released?
JG: The graphics, which is what matters most to high end users, would have made a real difference in winning market share. Plus, the number of publishers supporting the platform would have been on an entirely different scale. Who knows what might have been for both companies.
While 3dfx had adopted a multivendor approach for its Voodoo and Voodoo 2 cards to great market reach and success, it instead began manufacturing, marketing, and selling its own graphics cards with the Voodoo 3 onwards. What was the reason for this?
JG: Much like Apple, which delivers a superior user experience, doesn't need 10 companies bashing each other over the head in the market, neither did 3dfx. It's a simple strategy, build a great product and control the distribution process to ensure the highest quality and most positive end user experience.
In addition to the decision to start selling its own cards, 3dfx also had to contend with a graphics card arms race with competitor, NVIDIA. What do you feel enabled NVIDIA to quickly catch up to 3dfx and take over its market lead?
JG: Quite simply it was better management of their engineering resources and more engineers. It eventually got to the point where 3dfx needed two processors to stay ahead of NVIDIA's performance, and that priced 3dfx out of the PC OEM market, which was the bigger part of the business and ultimately led to the beginning of the end.
While the previous iterations of Voodoo cards were released in single generation gaps, 3dfx made the unusual move of releasing Voodoo 4 and 5 within the same time frame. What was the reasoning for this?
JG: Multiple factors including market strategy, OEM design cycles, competitor releases, 3dfx engineering cycles.
After seeing its market share disappear to NVIDIA, the company finally succumbed to bankruptcy in 2002. What was your reaction to the news of the closure?
JG: It was satisfying to see most of the engineering team find a new home at a great company and to be able to carry on their work. NVIDIA picked up a huge amount of engineering talent with the acquisition and some also went to ATI.
Were you surprised that 3dfx's assets were acquired by rival NVIDIA following the bankruptcy?
JG: No, it was good news and expected. It was a smart strategic move by NVIDIA and greatly accelerated its efforts in competing with ATI and eventually Intel. Also, 3dfx had received a favorable ruling in a patent case with NVIDIA, consolidation was happening throughout the industry, and it just made smart business sense for a lot of reasons.
While there has been much speculation about the real reasons for 3dfx's closure, what do you think were the main factors that contributed to it?
JG: The company missed design cycles and lost the engineering leadership battle to competitors who did a better job of delivering to their technology roadmaps. Look at the long history of the Intel versus AMD CPU battle, give all the credit you can to sales and marketing, but in the end it is the engineering teams that determine who will gain or cede market share.
Any truth to the rumors that 3dfx spent too lavishly on employee perks and lunches?
JG: I would dismiss this type of chatter. How much can you really spend on a lunch anyway? If you maintain a leadership position with your technology and you lead the market in revenue and profits, you can have anything you want for lunch.
What was your involvement with 3dfx's then next-generation "Rampage" project?
JG: Very limited.
Do you feel the "Rampage" project had the ability to save 3dfx had it come out earlier?
JG: There are a lot of "ifs" in this one, but yes. If it had come out much earlier, if it had performed much better, and if it had been a clearly superior solution, then yes, it could have saved the company. This is how fabless ASIC developers have always won or lost.
2012 marks a decade since 3dfx's bankruptcy. How do you feel seeing that the 3dfx brand and technology has all but disappeared from the PC space?
JG: All of the technology from that era is gone. The architecture has changed, the bus has changed, the memory has changed, everything has changed, and that is not unique to 3dfx. This happens with all technology in every category. Now mobile is emerging as a dominant gaming platform with an entirely new set of requirements. To its credit, NVIDIA anticipated this a long time ago and is once again incredibly well positioned.
Want to read other video game interviews with key figures from Sony, Microsoft and more? Then check out PC World's complete interview archive.