No Joy in Chipland
Moshe Gavrielov threw that delightful tidbit into his talk at this week's Semico Summit, an annual conference put on by analysts at Semico Research Corp. It's a good venue for learning about the state of the chip and chip-manufacturing industries, and -- because it's in Phoenix -- a nice opportunity to slip away for a couple of spring training games.
About that lack of VC money, though. Gavrielov's point is that it costs too much -- and takes too long -- to nurture a chip startup.
Yes, Xilinx stands to benefit if chip startups go away, because that would prod systems vendors into trying their own designs, driving up Field Programmable Gate Array (FPGA) demand. Gavrielov has brought up this point before, at Xilinx's analyst meeting in January, as evidence of the FPGA market expanding. (See Xilinx Minds the Gap.)
But I think his point is valid. Software tools have made chips easier to design, and foundries like Taiwan Semiconductor Manufacturing Co. (TSMC) (NYSE: TSM) mean you don't have to build the devices yourself -- but even so, the cost of making a cutting-edge chip will run in the tens of millions of dollars. Contrast that to the two-person "Web 2.0" operation that can toss out three ideas a week from Starbuck's (assuming they can wrest a table from all the laptop jockeys working on their screenplays). It might be unclear how those startups can make money, but it's a cheap gamble with potentially quick results, good or bad. Which one would you invest in?
The situation isn't permanent. Chips aren't going away. So, if no one invests in chips, then someone is going to start, because it would be an underfunded area. But what we won't get, at least not in the next cycle or two, is a bevy of chip companies all chasing the same target. The herd mentality just isn't there right now.
— Craig Matsumoto, West Coast Editor, Light Reading