Valley View: Vertical Integration is Back
Richard Irving, a partner at venture capital firm Pond Venture Partners, considers the purchase of PA Semi and Montalvo’s technology assets, by Apple and Sun respectively. The big surprises last week in Silicon Valley were Apple’s purchase of PA Semi, and Sun buying Montalvo’s technology assets. Why did these OEMs buy their own processor? How much did they pay and did that constitute a good exit for the entrepreneurs and investors? Until recently the conventional wisdom was that the x86 architecture reigned supreme in the computing space (all PCs and most servers), ARM-based processors led in mobile and low-power apps, and only comms infrastructure and industrial control remained CPU battlegrounds. Then Qualcomm launched Snapdragon and Intel its Atom, and the embedded space suddenly faced competition.
Apple’s PC shift to Intel made many think they would either use Atom for higher-end mobiles or continue with ARM-based chips for phones. Their move to acquire PA Semi is therefore surprising: with their excellence in software and industrial design, they would not seem to need their own CPU to win and maintain market share. But Steve Jobs has always liked proprietary hardware, so this may have been a key driver. And Dan Dobberpuhl’s ability to crank out low-power, high-performance CPUs of any architecture should not be underestimated. After all, he’s already done Alpha, ARM, MIPS, and Power. Sun’s purchase of Montalvo is a lot less significant. The term “asset purchase” means they picked up the pieces after the company blew up. Montalvo’s technology may appear as part of a future Sun core, but the acquisition does not suddenly give Sun a whole new level of CPU design capability. Reflecting this contrast in the prices paid, $278M for PA Semi will give the VCs modest multiples and should provide a decent reward for the employees, since the company was successful at raising venture funding at high valuations. Meanwhile after investing $100M in Montalvo, VCs will be taking away single-digit pennies on the dollar – not a happy ending. The lesson learned: starting a CPU company is a very tough investment proposition. Successes are modest and failures are very costly, likely leading most VCs to avoid them in future. So if you’re an entrepreneur, focus your startup on something other than building a new CPU.
See also: Electronics Weekly’s focus on x86 microprocessors, a roundup of content related to x86 microprocessor technologies and developments.
See also: Electronics Weekly’s focus on non-x86 microprocessors, a roundup of content on microprocessor technologies and developments not related to the x86 architecture (from ARM, Texas Instruments and MIPS).