I went to Lucio Lanza’s panel session on how much it matters that VC investment for fabless semiconductor companies and EDA companies has dried up. The rest of the panel was Sanjay Shrivastava from Denali, Gunjeet Baweja from Needham and Shishpal Rawat from Intel Capital.
And VC money has dried up. It now costs $50-100M to design a brand-new chip and get it to market and then, if you are lucky, you have a fabless semiconductor company that might get bought for $50-100M. The numbers just don’t work. It’s a similar story, but with lower numbers, for EDA companies. There hasn’t been a new EDA company funded by VCs for 3 or 4 years (just a few follow-on investments).
That huge $50-100M design cost is both terrible (nobody is funding chips) and an opportunity (it’s worth a lot to reduce it). It reminds me of the two shoe companies that sent their top shoe salesmen to Africa when shoemaking was first automated 100 years ago. The first writes back: “no opportunity here, they don’t wear shoes.” The other writes back: “enormous opportunity, they haven’t started wearing shoes yet.” Clearly one enormous opportunity in EDA is to reduce that $50-100M cost to $5-10M. To remove the $2M/month for a year cost of verification for a chip. To halve the design cycle.
There’s also the aspect that nobody knows what will happen to cause an upturn in the semiconductor customer base, the health of which is pretty much a prerequisite to a healthy EDA industry. As Lucio said, “Discontinuities are not forecast, they are only taken advantage of.” Nobody forecast the cell-phone market, Cisco, the PC, DVRs, video-games etc. But lots of money was made bringing entire ecosystems into existence to deliver them. Who knows what will be the next cell-phone?
The panel all seemed to think it was a good time to start a company. Clearly just as much innovation happens in a downturn, so in that sense it is true. Working in small innovative companies is always fun, but I think that there has also to be the chance of financial reward to make it attractive versus working for a big more secure company at a higher salary. Shishpal thought you should do it simply for love of innovation and technology, but the rest of the panel (and I suspect most of the audience) thought that this was unrealistic.
Or else, instead of the chance of making a significant capital gain we should all go and work in the public sector and have great retirement benefits anyway, paid for out of everyone else’s capital gains.
The bottom line is that the only way to really make an EDA company work today is to bootstrap it. Denali is proof that it is possible to build a full-scale company this way. More likely is to keep the company so small that even a $15M exit is attractive. But I question whether a company built this way is going to make an enormous dent in the cost of design. It is possible to develop a little piece of incremental technology this way that fits in the normal flow. I’m not sure it is possible for a handful of people to create technology that completely disrupts the way design is currently done.