Chip Trio Faces Post-Bubble Blues
A few years ago, all three were jewels of the semiconductor industry – telecom boutiques enjoying the high margins associated with long-haul networks. But each had to spend 2002, and much of 2001, coming up with a new game plan.
The results are now taking shape, as reflected in the companies' earnings reports during the past week. Vitesse's escape from long-haul is almost complete, although it's still got high hopes for Sonet chips. PMC hasn't fled the service-provider scene yet, but executives are shoring up other businesses as contingency plans. But AMCC has stood pat for now, although the company's got the cash to find a new direction if it has to.
Vitesse Stores up for Winter
In Vitesse's case, the changes made for a surprisingly upbeat call with analysts on Monday. The firm isn't quite storming back, but CEO Lou Tomasetta said Vitesse's makeover is starting to yield results (see Vitesse Reports Flat Q1).
Vitesse's long-haul sales are shrinking, but revenues in every other segment grew last quarter, Tomasetta told analysts. Storage revenues of $18.6 million were up 9 percent from the previous quarter, representing nearly half of Vitesse's revenues. Enterprise and metro sales likewise grew strongly, although both are still in the single-digit millions.
Tomasetta said the 24 percent growth in metro, to $6.6 million, was boosted by the introduction of the 9118 framer, which handles next-gen Sonet features such as virtual concatenation and generic framing procedure (see Vitesse Joins Next-Gen Sonet Party).
But it's storage, with roughly 10 percent growth per quarter, that has "been our strongest market for probably the last year," Tomasetta told analysts.
So far, "storage" has meant Fibre Channel chips, such as the 20-port device Vitesse developed with technology from Vixel Corp. (Nasdaq: VIXL; see Vitesse Ships 20-Port FC Chip). Vitesse is also entering the serial ATA market with a chip expected to produce revenues in the June quarter (see Vitesse Ships Serial ATA Chips), and the company's started backplane systems work as well.
Storage could be Vitesse's future, but it's still a bit unclear what a networking-storage-chip company would look like. "There's the whole other side of storage, which is more PC-ish" and carries the stigma of low margins and tough competition, says Sandy Harrison, analyst with Banc of America Securities LLC.
That's a particular concern when it comes to the serial ATA chip. Tomasetta maintains that Vitesse is aimed only at the higher-margin server market, while the competition focuses on PCs, but it's not clear whether server and PC chips will remain that different in this area. "For them to pull 60 or 70 percent margins in serial ATA would be a real accomplishment," Harrison says.
PMC Goes Micro
Like Vitesse, PMC-Sierra is looking at storage (see PMC-Sierra Reports Q4). It's offering Fibre Channel silicon and is aiming its serializer-deserialzer (SerDes) marketing towards storage customers. But the ace up PMC's sleeve could be its microprocessor division, created with the acquisition of Quantum Effect Design Inc. (see PMC-Sierra to Buy Quantum for $2.3B). Its MIPS-based processor can be the control-plane brains of a switch or router, but 60 percent of its business has been in printers, and another 30 percent in various enterprise boxes.
It's a promising division, but it represents a shift in PMC's philosophy. "We've been more adept at understanding lower-volume applications at a higher margin, rather than the lower-margin, higher-volume products," says Doug Brownridge, vice president of corporate marketing.
PMC responded to the downturn rather quickly, but its early layoffs didn't prevent another round of cuts at the end of 2002 (see More Cuts at PMC-Sierra). "There's no question the two companies who moved first were JDS Uniphase Corp. [Nasdaq: JDSU; Toronto: JDU] and PMC," says Jeremy Bunting, analyst with Thomas Weisel Partners. "But as we enter 2003, telecom continues to be moribund."
The layoffs were a surprise, considering PMC had taken most of its lumps much earlier in the downturn. But the overall direction jibes with the plan PMC outlined in 2001 (see PMC-Sierra Pulls Packet Silicon). The major difference is that the company is moving even further than expected from the network core, into enterprise and storage.
"It's not that we're turning away from service providers. We continue to play a dominant role there," Brownridge says.
PMC officials also hinted at products in the 10-Gbit/s transceiver space, the realm of the Xenpak, XPAK, X2, and XFP multiservice agreements (see Is Xenpak Past It? and The X-Wars: Agilent Strikes First). They wouldn't give specifics on their plans, however.
AMCC Digs In
Finally, AMCC has stuck to its guns in telecom and network processing (see AMCC Reports Q3). Company officials continue to hint that they might use acquisitions to diversify – which makes sense given AMCC's cash reserves of $1 billion (see AMCC May Go Shopping).
But don't expect AMCC's makeover to start any time soon, Banc of America's Harrison warns. He doesn't think AMCC is "complacent," but he also expects that chairman and CEO Dave Rickey wants to make sure the company doesn't rush blindly into some new, temporary fad. "Once we find out which direction the industry's going in, then Dave Rickey will use his $1 billion," he says.
And, like PMC, AMCC might not be done cutting costs.
"They've had the attitude that they can finance their way through it, but investors these days are looking for a return to profitability," says Weisel's Bunting. "There's no question that for them to return to profitability in, say, a year, they will have to cut more."
— Craig Matsumoto, Senior Editor, Light Reading