Valley Wonk: Chips Are Down
Not that you can tell by the quarterly earnings reports. As we've noted already, chip folks are chipper, telling investors this will all blow over soon (see Chip Gloom Shall Pass). Agere Systems Inc. (NYSE: AGR.A) was the latest, with CEO John Dickson saying last Tuesday that he sees inventory problems in enterprise and telecom extending "some way through the first calendar quarter of 2005." Dickson cast that as glum news, but other executives have used the same target – early 2005 – as a good sign, indicating the problem will be shrugged off soon.
Jim Feldhan, president of Semico Research Corp., has his doubts. His firm's numbers show the problem lingering through all of 2005, and I'm inclined to side with him.
You have to realize that chip executives have a track record of being fabulously wrong about this sort of thing. It's like having a contractor work on your house. Everything will get done in a few months, you'll see. A year later, the kitchen still has no roof and you're wondering if you'll ever get your indoor plumbing back.
Semiconductor demand moves in cycles, and the "up" cycles usually end because of an inventory overflow. Where companies tend to mess up is in predicting when the overflow will happen. If they slow down production too late – as it appears they've done here – the overflow gets too large, and the resulting "down" cycle becomes longer.
That's how the boom of 1995 ended. At the time, the Internet was just beginning to take off; personal computers were invading the voluminous home market; and multimedia (back when video and sound were special enough to earn their own buzzword) was "adding value" to the increasingly commoditized PC. The Comdex tradeshow mattered. The tech economy was booming, and that led the semiconductor industry to declare semiconductor cycles dead.
Dead! No more cycles! And it wasn't just because things were going well; it was because after all these decades, executives had become smart enough not to let it happen again.
About four years later, I got the chance to interview Mike Hackworth, then the CEO of Cirrus Logic Inc. (Nasdaq: CRUS), a "multimedia" high flier. And Hackworth said it was at that point -- when CEOs started giving headline quotes about the death of semiconductor cycles -- that he knew the boom was dead. In his gut, Hackworth realized the market loves to punish that kind of hubris.
He rode the wave anyway, and Cirrus lost out. By 1997, currency destabilization had socked Indonesia and Korea, and the economic shockwaves were enough to bring down the tech boom.
So, the era of "cycles are dead" created the worst cycle executives had seen to date. One executive after another told me, "I know we say this every time, but in my 20- (or 25- or 30-) year career, this is the worst it's ever been."
This year's chip slump won't be the worst ever, but it's not going to dissipate quickly unless we all chomp ASICs instead of M&Ms this Halloween. One or two chip companies having problems would be OK, but it seems every company is drowning in inventory.
And it's not just the semiconductor firms. Distributor Arrow Electronics Inc. (NYSE: ARW) reported inventory overhangs, too. That means one of the biggest buyers of chips doesn't need to buy for a while. Worse, on Tuesday, foundry Taiwan Semiconductor Manufacturing Company (TSMC) (NYSE: TSM) said it's expecting "almost flat" growth in 2005 – this from the largest manufacturer serving the chip industry.
It's possible that all these factors could correct themselves in one or two quarters, but that would take a massive synchronization effort among lots of large companies. That's not going to happen. And even if demand kicks into gear in 2005 – and Feldhan doubts it will – we'd have to rely on the industry to show enough restraint to prevent yet another inventory overhang and another down cycle.
Feldhan's analysis is looking good: The tipsiness of early 2004 is going to create a hangover that lasts for most of 2005.
We'll be delving into this issue a bit more next week, but the long and short of it is that anybody counting on a quick return to explosive growth in the chip industry is being misled.
— Craig Matsumoto, Senior Editor, Light Reading