Has anyone else noticed the mass exodus by high-performance analog semiconductor manufacturers from the power management IC (PMIC) market? Last year Gregg Lowe let the cat out of the bag in a subtle way when he outlined his PMIC-less vision as the new CEO for Freescale. Lowe said it was a flawed investment for Freescale since PMICs are difficult to design.
If those weren't words from the former top analog guy at TI, the world's leader in power management, maybe I wouldn't have had a second thought. But frankly, I don't buy the claim of “hard to design.” I think it is proving easy and cheap.
Lest one think Lowe's perspective pertains only to Freescale, consider Intersil. Intersil’s May 2012 Investor Analyst Day presentation included a slide 77 that distanced that company's past and present in power from a future focused on data converters, high-speed amplifiers, and temperature sensors. And then there is Maxim.
In 2010, Maxim was touted by Intel as one of three PMIC suppliers for its then upcoming Atom processor. Maxim's CEO, Tunc Doluca, in early 2011 wrote an EETimes article touting the fortunes waiting to be made in portable power. But then last week Intel reported Dialog Semiconductor is now a preferred PMIC vendor for its next-generation application processor. No mention of Maxim.
So what is Tunc-and-Team betting on now? Well, last November he was talking with EETimes about Zeus with analog frontends, data converters, and Big-D, little-a. I had to reread that little-a part a few times to make sure this was the same house that Jack built.
If those three examples don't portend the mass exodus, look at what's happening at Apple Computer.
Apple is one of the world's largest purchasers of power management products. Curiously though, I've not seen a single PMIC product from any of the top players show up in any Apple teardown reports. But Dialog Semiconductor is front and center.
Dialog's home page prominently proclaims “The power to be…” No one can doubt its claims of power, focus, and expertise, considering Dialog’s 2012 sales of $744 million will represent a nearly tenfold increase over a short five years. This is astounding since Dialog's 47 percent year-on-year growth coincides with a period when the worldwide revenue for the other PMIC powerhouses is flat to down.
So how did Dialog do this in just a couple of years with two orders of magnitude fewer patents and no internal specialized analog fab, assembly, or test? Simple. In the minds of 100 million people, its products are good enough, and they are sold to Apple for a very nice price.
How nice? Well, if Dialog's gross margins are any indication, probably less than 38 percent nice. Compared to the old guard with typically 60 percent gross margins, this is more like Christmas than a high-performance analog business with high entry barriers. Let's be honest here. If any of the old-school PMIC powerhouses sold their products at 38 percent gross margin investors, would flee their stock like it was Enron.
Look, when a hard-to-design product that used to sell at nearly 3x manufacturing cost now sells for about 1.6x, it loses the label of being hard to design and becomes a commodity available from the low-cost leader. Think Wal-Mart in the US or Metro in Germany: great companies, but not mistaken for Tiffany's.
I think the race to data converters will be short-lived as well. The fact that Cirrus Logic, another Apple Computer fanboy, can sell its data converters at over 2x cost is misleading. Two weeks ago, Dialog announced another PMIC design win with Samsung, and that one includes — you guessed it — audio data converters.
The clock on hiding behind high-performance analog patents is quickly running out. The innovation asymptote is rapidly approaching. It's time to start moving up the food chain vis à vis Maxim's Big-D, little-a strategy.
Oh! And what about Rich Beyer, the former CEO of Freescale and Intersil? Last week we learned he is a new director at [drum roll, please] Dialog.