Issue: Volume: 24 Issue: 8 (August 2001)

Turning Point

In the past two years, nothing seemed to stand still or remain the same. SGI abandoned graphics chip development for the desktop and worked out an arrangement with Nvidia for workstation graphics. Intense3D was acquired by 3Dlabs. Nvidia acquired the software team and OEM clients from Elsa. SGI acquired Intergraph's workstation group. Evans & Sutherland pulled back from developing workstation AIBs and chips. HP reorganized and consolidated its workstation divisions. NEC entered the market with a powerful AIB. ATI acquired the FireGL division of SonicBlue (formally S3). Three innovative chip design groups (PixelFusion, Raycer, and Philips' SP3D) pulled out without completing their designs (Apple picked up the Raycer team, PixelFusion moved to communications). SGI enhanced its Octane systems. Sun introduced a low-cost range of products. And HP enhanced its fx graphics.

The net result of all the assimilations, acquisitions, and mergers, was a 29 percent drop in the total number of suppliers of workstation graphics AIBs and chips during 1999, followed by a period of equilibrium during the past 18 months, when as many suppliers entered the market as dropped out. However, the recent stability is like the quiet before the storm. Given the imbalance in the number of suppliers per units sold and per revenues, especially compared to the ratios in the PC market, there will be more casualties.

The primary driver of this upheaval will be the continuing migration of low-cost, reasonable-performance hardware from the PC market, which is squeezing the profitability from the business models of professional graphics hardware vendors. The migration of low-end graphics into the workstation market has largely taken place in the Windows NT/2000 arena, at the low-end and mid-range. And market demand for this class of workstation has been healthy. In terms of shipments of Windows NT/2000-based workstations (and accompanying add-in graphics boards), the market exhibited a compound average growth rate (CAGR) of 48 percent from 1998 to 2000 and is predicted to have a 14 percent CAGR from 2000 to 2006, resulting in an overall CAGR of 22 percent from 1998 to 2006.
Sales of Windows NT/2000 workstations (including add-in boards) will grow 379 percent from 1998 to 2006.

At the same time, the average selling price of workstation graphics boards has declined 52 percent from 1998 to 2006. However, unit sales will in crease 379 percent from 1.53 million units in 1998 to 7.31 million in 2006, creating overall market value growth.

One cause for concern is that although the market will expand steadily in the next few years, growth will fall off after 2003. Indeed, with the ever increasing power of processors and graphics controllers, combined with the lackadaisical employment of graphics board acceleration features by software vendors, users will question just how much productivity they will get from new systems. Therefore, we predict a year or two of confusion and some overbuying, as users experiment with various combinations of 32-bit super high-performance AIBs, and the new 64-bit AIBs to see which combination works best for their applications.

There's good news and bad news in this equation. For the box and AIB builders, there's good news because there will be over-purchasing of systems, which will give a temporary spike to the market, but bad news because users will then pause in their replacement buying while they figure out what works and try to extract the return on investment from the machines. For buyers, there's good news because as they pull back, the box and AIB builders will react by lowering prices, but bad news because if the turmoil decimates the ranks of workstation AIB suppliers, buyers' choices will be reduced to only two or three companies.

Jon Peddie is president of JPA, a market research and consulting firm that serves the computer graphics industry. He was recently named one of the top 10 most influential industry analysts by Adweek magazine. He can be reached at

Jon Peddie