In-Stat, the venerable market-research firm born in the big heart and tireless tenacity of an ex-Motorolan, died in its sleep this week. The cause of death was old age and market myopia.
Founded a generation ago by Jack Beedle, a man who literally and figuratively was larger than life, In-Stat became for years the industry's go-to place for market analysis. It niched the much larger Gartner/Dataquest by focusing relentlessly on semiconductor market data.
Beedle hired brilliant young engineers and analysts and worked them like dogs; many, while thankful for the opportunity and early tutelage, chafed at Jack's style. After gaining street cred under Beedle, many went off and started their own successful businesses.
Eventually though, Beedle aged, his health was failing and he sold the company to Reed Business, which kept In-Stat going for a decade. It was never the same without him, and he died not long after at 73.
Desert legends The In-Stat desert gatherings were ostensibly an opportunity for industry execs and inked-stained wretches (yes, it was all print then) to gather and listen to forecasts and presentations from industry leaders. But the real action was in the hallways at coffee breaks, where you watched carefully to see which execs were huddled in which corners, possibly hatching merger plans or a new foundry deal.
Late at night, it got really entertaining, when the official festivities closed and attendees staggered off to a resort bar to keep on keepin' on. I remember one evening cornering a CFO I'd quoted in an old story. He gave me the "real scoop" that night, but, in the din of the crowd, I lost any valuable nuggets he gave me because, while he spoke English, he was Scottish. And we had no translator.
But it was those far-corner huddles where mergers were afoot that ultimately undid market research firms. The industry consolidated. IC companies got huge and took that stuff in house. Their gaze turned toward financial-research houses and their sponsorships and interest dried up or were shifted to their own "walled-garden" events.
So be it. It's the evolution of an industry.
But the hollowing out of many of the service components of the electronics industries (and, in In-Stat's case, the immediate cashiering of 30 analysts) has left us nearly impoverished. There are fewer filters, fewer checks and balances.
"We know our customers" is the usual retort. Yeah. Sure you do. It's not knowing your customers that you should be worrying about. It's figuring out who your customers in five years should be, and you can't forecast that with institutional hubris.
That filter function, that early-warning system is where analysts made their bacon, back in the day, back when outsized figures like Jack Beedle roamed the earth.
It would be great fun to create a list of some of the other colorful journalists and analysts who were active in producing semiconductor news (and gossip) in the 1955 to 1990 period of ebullient semiconductor growth. Don Hoffler comes to mind as do some of the news hounds like Marty at Electronics News. Sometimes we got very upset with them but they sure kept things stirred up.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.