editorial

What's a Reasonable Price for EDA Tools? The practice of wringing the greatest possible discounts out of an EDA vendor may backfire over the long term by stifling investment and innovation. by Tets Maniwa

As the opening of the Design Automation Conference looms before us, we anticipate various offerings from the EDA community. Many of the DAC sessions focus on current research in EDA topics and the development and implementation of the latest technologies and tools. Academic and corporate researchers will describe their efforts in overcoming the hurdles that IC manufacturers present to the design community: e.g., the ever-increasing circuit densities within ICs and techniques for accurate and complete circuit design.

All of this research costs money--and lots of it. In fact, EDA vendors spend over 25 percent of their revenue on research and development. They must develop new algorithms for the tools, validate the functions, incorporate the functions into the underlying tool structure, and create interfaces for users. Ancillary utilities enable the tools to read and write data to other tools. The feature set for a new tool costs millions of dollars and can take as long as three years to develop and commercialize. To fund this development, the start-ups burn their venture capital and the established companies depend upon their current revenue. But even though such EDA vendors spend a good percentage of their revenue creating software in order to solve design problems, users are trying to minimize the amounts they pay for those tools.

The practice of aggressively negotiating very large discounts from the list prices causes the EDA vendors to have lower revenue and longer payback periods for their R&D investments. Longer payback periods--in financial terms, lower returns on investment--depress the company's stock price, making it harder for the EDA vendor to raise money from other sources. This in turn further restricts potential funds for R&D or makes those funds more expensive by increasing the interest rates for any loans. Subsequently, the lower valuation of the whole industry makes the job of starting up a new company harder, since the rewards for the investor are smaller than they would be for alternative industry investments. But a more direct result of the reduction in revenue is the reduction in the money available for engineering. Everyone knows what happens when engineering is cut; the work takes longer to complete or the product may ship with some unaddressed bugs. A reduction in resources, however, doesn't excuse EDA vendors from shipping tools with bugs at any price.

The EDA industry is just like any other manufacturing industry, except that the companies' resources are mainly tied up in brainpower rather than in physical equipment. One difficulty the EDA companies face is the relatively small size of the market. The number of potential licenses for any tool category varies from hundreds in the very big back-end physical design tools to tens of thousands for simulators and FPGA tool suites. The maximum potential revenue is the product of the average selling price and the total number of units. To make a million dollars, you can sell 3 or 4 copies of a place-and-route tool or 50 to 500 copies of a simulator. In a restricted market, where the number of customers is fairly small, decreasing the price doesn't increase the product volume; the lower price just leads to lower revenue for the EDA vendor.

Although the cost of software delivery is very low--a few dollars for CDs--the true value of the software is a function of the underlying abilities of the tool. The EDA industry-developed tools have increased designer productivity by many orders of magnitude. Who today would even consider hand drawing waveforms or generating transistor-level netlists? A fair price for EDA tools should be in units of engineer equivalents. If a tool increases a designer's productivity by a factor of 100 over doing the task manually, then the tool should be valued at 99 engineer equivalents. Most of today's tools cost far less.

By driving down the price of essential tools, the user community is endangering future innovations and the EDA companies' ability to address problems that the next generation of silicon will engender. Without sufficient funding for more R&D and reasonable expectations for a good return on investment, where else can we look for tools to design the multimillion-gate systems on a chip in the next century?

Please let me know what you think by sending your comments to tets@isdmag.com .

To voice an opinion on this or any Integrated System Design article, please email your message to jeff@isdmag.com.

integrated system design  June 1999