Some EDA tools are too complex for average engineers to use due to the complexities of leading edge semiconductors fabrication processes. Vendors need to adapt their business strategies to this new market environment.
In the last two weeks I heard some things that make me worry about the capability of EDA to grow its revenue much larger than it is today. May be we can achieve $5 billion, but drastic changes seem required to go much beyond this figure.
First, Open Silicon has purchased the IP assets of Zenasis Technology as the latter closed its doors. In commenting on the acquisition, Naveed Sherwani, Open Silicon president and CEO stated that Zenasis technology did not really belong in an EDA business model because: "The technology is clearly sophisticated and has a long history of academic research. But what the technology does not have is ease of use. The point is, this is not something an average engineer could use right away, which is a critical need for a standard EDA tool."
Second, in a presentation to editors and analysts by Ponte Solutions, Jim Hogan showed that only five consortia of semiconductors vendors and less than a handful of independents would be viable in providing 45nm wafers. The company has an interesting business strategy that targets both those companies and their potential customers, yet it may only have no more than 60 companies as its total available market.
It is a fact that complexity is one of the main reasons for the high costs of product development at 65nm, and it is also undeniable that only a minority of system houses are targeting this process node for their products today.
I think that teams doing 65nm designs are not staffed by average engineers, and those that will tackle 45nm designs will be even less average. So, Zenasis probably had about half the total available market (TAM) that Ponte has, since both target semiconductor companies as their customers. Zenasis had only twelve persons on its staff, so can Ponte sustain a staff of more than thirty or so and still be profitable? Clearly the EDA market is divided into at least two segments: those doing designs for leading edge processes, and those that do not. The latter is a segment destined to grow at a much greater rate than the former. And most of the growth will occur outside of the US since staffing costs are less in developing countries and "only" average engineers are needed for those jobs.
The electronic industry has often reassured the government and the professional engineers that leading edge designs will remain in the US and thus "offshoring" is a phenomenon not to be afraid of. But now a new message is emerging. Leading edge markets are small and characterized by strong and stable business connections. Tools will have to be more specific to a given foundry consortium than generic, since generalization increases complexity of use.
EDA companies may want to look closely to the PCB market, where there is a marked difference between tools for leading edge designs and those for still demanding but less complex PCB designs. It is also clear that leading edge designers will require a closer technical cooperation from EDA companies than the other class of customers. Cadence has already started to implement such a vision. It has tiered its products portfolio and introduced the Kit line of products that includes consulting services as an integral part of the product. But we still do not know how the revenue streams will behave once the new market profile takes hold: will EDA grow or become a service sector with o particular attraction for investors? And what will happen to all those engineers not lucky enough to be "above average"?