Fishman, at ADI since 1971, has seen many downturns and upturns -- perhaps, more than he cared to see. The semiconductor business, after all, is cyclical. But even among chip-biz veterans as grizzled as Fishman has, the cycle is tough to nail down. "We can never quite predict when the recession is coming," said Fishman. "The only thing we can predict is that the recession will happen."
So, how does a company prepare for a recession whose timing is a mystery?
Fishman suggested: "Be responsive to the reality. Recognize it and don't fight it. Be responsible on costs.'"
As simple as it may sound (and it comes down to common sense), these are tough rules to stick to.
Every company has businesses -- or pet projects -- from which they never expect a return on the R&D investment. Fishman, known for his bluntness, made it clear that for those underperformers at ADI, "we've decreased investment" or simply cut them off.
The case in point is the 2007 sale of ADI's baseband chip product lines and some of ADI's cellular handset baseband support operations to MediaTek, Taiwan's fabless chip company. One could argue that that hard choice by Fishman might have been the key to saving ADI from disaster during the global recession a year later.
The reprioritization of ADI's business is happening on a much broader scale.
One no longer hears much about its DSP business.
Fishman said, "We still remain an analog company." The reason ADI remains in the DSP business is its customers, who demand integrated solutions which require more signal processing functions, he explained.
That, however, is a very different picture of ADI from the image promoted very heavily by the company as an entity possessed with two major product/technology pillars: analog and DSP.
It's no secret that ADI today has pretty much cast its lot with Blackfin DSPs. Some in the market also appear to believe that the company has stopped developing its Sharc DSP lineup. Fishman, however, insisted that the company is still investing in Sharc. He noted that Sharc still dominates the floating-point DSP market, and there remains a huge opportunity for its applications.
But when asked about the company's investment plans -- specific to Sharc, Fishman said, "Our investment is going into derivatives [of Sharc] to offer specialized functions to our customers. Will we invest in the next big cores? No." Further, Fishman elaborated, "Rather than investing in the next big core, our focus is to increase system performance through accelerator functions to better leverage the exisiting code base."