EDA software: complex...and expensive
A report published in Bloomberg this week reveals that of more than 50 of the top high tech companies based in Silicon Valley, two of the EDA industry leaders – Cadence and Synopsys – are at the top of the list when it comes to percentage of revenue spent on research and development. Cadence was at the very top with its 40% (or almost $350 million) investment in developing the new tools that chip designers need to keep pace with innovate new products. Synopsys wasn’t far behind at #3 with its 32% spend (and even larger absolute amount: $433 million) on R&D.
The eye-opening statistic, though, is that the average R&D investment by tech firms hovers around 10% – with software companies averaging around 11.4 % and hardware companies at 7%. The huge gap between EDA R&D spending and the rest of high tech (interestingly game maker Electronic Arts was in the same range as Cadence and Synopsys at 34%) tells me one of two things: Either EDA is an incredibly expensive technology to develop, or EDA companies are not getting enough for their investment in new products. Or both. The latter has always been the argument in EDA, as executives and analysts have long pointed to to the discrepancy in the size of the market EDA serves compared to the relatively puny $4 billion in sales that EDA tools and services generate.
Personally, I don’t see any way around the situation. Despite what Cadence may want you to believe with its EDA 360 mantra, chip design is finite industry with a pretty fixed number of potential customers. The number of customers is not growing (in fact, it’s probably shrinking, or at least moving to lower cost models through off shoring design activities). Unlike the markets that EA serves, there aren’t alot of new opportunities emerging to sell new EDA products to – and the installed base is essentially a replacement business. And, customers have a long-ingrained sense of what they should be paying for EDA.
On the flip side, developing EDA technology is obviously a complex – and expensive – process. It’s difficult to lower the cost model because of the caliber of people required to build state of the art EDA tools. So spending big on R&D is a necessary part of the game in EDA, and the industry spends less on other functions, like marketing, for example, because of the very focused markets it serves.
Which is also why more consumer oriented companies spend a lot less as a percentage on R&D. Apple was near the bottom of the Bloomberg list, with just 2.9% spent on R&D.
