Today’s guest blog is from Isadore Katz, CEO of CLK Design Automation. He has been in EDA for 20 years at Chrysalis, Cadence, Metasoft (Hspice) and even escaping EDA for a time as CEO of Lightchip before the gravitational pull of CLK brought him back in.
Four Companies Enter, Three Companies Leave
The current thinking these days is that anyone attempting an EDA start-up is an idiot. The VC’s are uninterested, exits are few and far between, and the chance of an IPO is zero. As the CEO of an EDA start-up company, I guess that makes me Idiot in Chief. So take everything I am about to say with a grain of salt; say the size of Grand Cooley Blog.
There are too many IC tool EDA companies. Not start-ups, there are always too many start-ups, but big IC EDA suppliers. Two is company, three’s a crowd, and four – four is just bad news. One of them has to go.
Both the semiconductor industry and EDA have to restructure in order to survive. The semiconductor houses must establish some level of pricing power with their customers. The world does not need 10 IC shops selling weakly differentiated wireless components to four handset manufacturers. Likewise the EDA industry needs to regain leverage. We do not need four major vendors supplying strained, front-to-back flows that get discounted to the point of oblivion.
Today all four EDA majors are trying to compete for the same business from the same 30 semiconductor shops, all of which are under severe margin pressure themselves. The result is an unmitigated disaster for all concerned.
Do the math. If four companies compete for 30 companies’ business, that averages out to 7.5 wins each. If 3 companies compete, its 10 wins each. More potential wins per major player means less discounting/higher revenue per customer, more top line revenue/bottom line profits and growth per major EDA vendor, and better industry performance. All of which should improve our image with investors, which ultimately should translate into improved market capitalization for all.
Why should the semiconductor companies care? Because the accountants’ persistent belief that EDA is a tax, and lots of inadequate, but heavily discounted, tools is a deal, is simply wrong. The real cost culprit is the difficulty of getting working silicon out of the tools. It is supposed to be electronic design automation, not 100 iteration, 50,000 lines of scripts per tool, sort-of-kind-of automation.
EDA’s job is to make the world’s most complicated version of Photoshop — including the camera/capture tools (but not the printer). There are at least 50 unique components to this flow. None of the big four does them all well. None of them can expect to maintain the level of innovation required to keep each of these at the leading edge needed for the latest process node. In fact, most of them have massive trailing edge commitments to customers that exist only as small font bullets in an archived PowerPoint presentation. Worse yet, some of them have a stunning trail of press releases promoting imaginary products. The current state of 4 major vendors with diminished pricing leverage and market caps will only perpetuate this condition.
With 3 healthier major EDA players, with market caps that reflect their revalue, focusing on what they do well, acquiring what they need, and making sure that producing 40nm and 32nm becomes practical and not just possible, maybe we have an industry.
Of course, I could be wrong. Even this may not be enough. In which case look for my future rant, “Three Companies enter, Two Companies leave.”