The electronic design automation industry is poised for rapid growth in 2004 and 2005 because of the recovering IC market and the increase in R&D spending by semiconductor companies. Significant challenges however, will make achieving that growth a daunting task both now and in the future.
The Semiconductor Industry Association predicts that the 2004 semiconductor market will approach $200 billion, representing close to 20 percent growth over 2003. The EDA market, on the other hand, is less than a $4 billion market that will struggle to grow by 5 percent in 2004, up from 0.1 percent in 2003.
The conventional wisdom in the IC market is that IC design completions in the coming years will be flat or down, but design implementation costs will continue to increase as complexity grows.
Hence, the revenue per design will need to increase to cover the increased costs of design implementation. The latest estimates from International Business Strategies, a research firm based in Los Gatos, Calif., indicate that the expected number of design completions will range from 6,400 to 6,800 per year over the next four years, with only 12 percent of those designs-fewer than 800-being at 90-nanometer and smaller process geometries by 2007. At an estimated $50 million total cost per 90-nm design, only high-volume applications, such as consumer and wireless, will be able to afford to go there anytime soon.
The reason for such slow growth expectations in the EDA industry is that since design activity continues to be heavily focused at 130 nm and above, there's a reduced need to procure new EDA flows to handle the increased parasitic effects of 90 nm and beyond. Companies will continue to pressure their engineering organizations to utilize their existing tool flows, to add tool capacity only when essential.
Unfortunately, to be competitive, the top EDA vendors will not only have to continue spending R&D dollars and acquiring EDA companies to address the tool
and methodology issues at these advanced nodes, they will also be forced to examine additional ways to get a bigger piece of the growing semiconductor market pie.
Semiconductor IP move
One way will be to adopt more of a value-based pricing model. The other will be to attempt to grow into other related markets, most notably semiconductor intellectual property (IP).
As feature dimensions shrink, EDA tool costs are declining as a percentage of design project costs. Even though budgets for design implementation are increasing within semiconductor companies, the IC industry needs to better understand the value of EDA tools in order to allocate larger budgets for them. This will be difficult for the EDA vendors to pull off, however, because they've conditioned the market so well over the years with "all you can eat" fixed-term licensing models.
The second major opportunity for fast growth in the EDA industry is through entry into related markets. One such target for EDA companies is semiconductor IP. Semico Research forecasts the IP market to grow around 29 percent in 2003 and 35 percent in 2004, reaching close to $1.5 billion in revenue. IP companies have found ways to get a faster-growing piece of the semiconductor pie through such arrangements as royalties.
For future growth, the EDA industry needs to become more closely coupled to the financial metrics within the IC industry than in the past. As it looks for answers, EDA companies are realizing that they can learn a lot from the semiconductor IP industry.
Jim Ensell is vice president of marketing at Virage Logic Corp. (Fremont, Calif.).