Stardent and the Elusive Graphics Supercomputer 333
and product costs. However, the real flaw was introducing a first product that failed to demonstrate clear performance superiority, as discussed in Chapter 8. Although a second set of products overcame this shortcoming, the inadequacy of Ardents first product inflicted very painful damage on the firm. According to Allen Michels, Ardent's founder and CEO: "You get one shot."
Rarely does a start-up get a second chance if it misses with its first product. The first product must be designed, sold, and delivered correctly. The only reason the company survived to merge with Stellar in October 1989 was that Ardent's Japanese partners had a basic faith in the firm and the market. Poduska became president and Michels chairman, and in 1990, Michels and Sanders left the organization.
The following are some lessons that can be learned from Ardent's experience with its graphics supercomputer:
Stardent underestimated the competitiveness of Silicon Graphics by classifying it with established companies such as Apollo, DEC, HP, and IBM.
Ordinary venture financing would probably have given up on Ardent after three years. Kubota is determined to be successful in technical computing and is prepared to invest for a long-term gain.
Detailed planning is essential. With a proper seed stage, Ardent would have either cut the product to fit the funding or made a realistic product plan.