Optical components

Xanoptix Hatches Hybrid Strategy

Trying times in the optical components business are forcing companies to zig and zag. Startup Xanoptix Inc. today announced the hiring of a new CEO, part of a move to retrofit the company's manufacturing methods to the semiconductor industry.

That's not to say Xanoptix is changing its product plans. In fact, new CEO Rob Baxter says Xanoptix's parallel optics module is drawing sample revenues and should soon hit general availability. Baxter claims the best way to expand from there is to start offering the company's brand of hybrid integration -- a 3D stacking of finished chips -- as a service.

The service would target the traditional chipmakers, but it could also be offered to competitors in optoelectronics. "This technology should, if developed properly, get us into all aspects of the semiconductor business," Baxter says.

Baxter sold these kinds of services as a senior vice president at Chartered Semiconductor Manufacturing Ltd., a Singapore-based contract manufacturer that competes with the likes of Taiwan Semiconductor Manufacturing Company (TSMC) (NYSE: TSM). His hiring is being announced today, with prior CEO James Norrod stepping aside but continuing on as chairman of the board. Xanoptix is also announcing it's added Russ Johnsen, former Analog Devices Inc. (NYSE: ADI) executive and another semiconductor veteran, to its board.

Here's why hybrid integration is key. Most chips are made of silicon and produced using a CMOS (complementary metal-oxide semiconductor) process. CMOS chips are relatively cheap to build, and multiple CMOS devices can usually be integrated together to save space and energy. But lasers and very high-speed components still have to be built from materials such as indium phosphide (InP), silicon germanium (SiGe), or gallium arsenide (GaAs).

There's no economical way to build a chip combining one of these materials and CMOS, a task sometimes called monolithic integration. The next best thing is hybrid integration: building a device in CMOS and a device in, say, GaAs, and connecting the two together onto a module. While monolithic integration requires sophisticated materials expertise, hybrid integration boils down to an assembly problem.

Xanoptix's core technology is a bit unusual -- it's a hybrid-integration process that stacks components vertically rather than laying them side-by-side (see Xanoptix Unveils Hybrid IC Technology). This technology is at the heart of Xanoptix's first product, a 72-transceiver parallel optics module called the XTM-72.

Baxter's plan, then, is to become a "stacking" foundry for chipmakers, receiving chips and building a hybrid-integrated device out of them. The idea is to introduce the technique to the greater semiconductor community. "We'll be looking at options of building products not in the optoelectronics space," Baxter says. "The thought process is that the capability we have in Merrimack [Xanoptix's New Hampshire home] will be available to other companies."

Baxter claims his company is in talks with several potential customers. He wouldn't give details, but a previous EE Times report noted that Velio Communications Inc. and Vitesse Semiconductor Corp. (Nasdaq: VTSS) have done joint research with Xanoptix.

What about Xanoptix's actual product, the XTM-72? Baxter says it's "ramping up" and generating sampling revenues, based on next-generation systems that are in development. "Probably in the next 12 months, you'll see customer shipments of products that use parallel optics," says Harald Hamster, vice president of business development.

Xanoptix was founded in 2000 with hopes of tapping an enormous parallel-optics market (see Integrated Optics: $2.6 Billion by 2005?). The recession postponed the need for 72-channel connections, but Baxter denies that the foundry strategy was developed as a contingency plan in case the transceiver business continues to lag. "There's a lot going on [in transceivers]," he says. "Please don't think that it's a poor business or a dying business -- we just want to expand beyond that."

A similar startup called TeraConnect Inc. is also developing hybrid-integrated devices (see Parallel Optics Boosts Bandwidth). One theory says Xanoptix was created by venture capitalists who'd been shut out of the TeraConnect deal, although Xanoptix has denied this (see Xanoptix's Strange Story). Another potential competitor is Sparkolor, acquired by Intel Corp. (Nasdaq: INTC) in September 2002 (see Intel Acquires Sparkolor).

Xanoptix has about 50 employees. The company has raised $80 million, half of it in a January 2002 funding round (see Xanoptix Lands Cash, Launches Product).

— Craig Matsumoto, Senior Editor, Light Reading

Sign In