Cypress Flexes MEMS Muscles
It was, but it’s turning out to be more significant than it first appeared. Silicon Light Machines has a totally different take on MEMS technology from other manufacturers targeting telecom applications. The upshot is that it can make a wider range of components -- everything from variable optical attenuators (VOAs) and tunable filters to dispersion compensators and ultra-fast optical switches.
Silicon Light Machines was founded in 1994 by David Bloom, a professor at Stanford University, where the original idea of “Grating Light Valves” originated. Its board of directors included Andy Bechtolsheim, a founder of Sun Microsystems Inc. (Nasdaq: SUNW) and Granite, an early Ethernet switch startup acquired by Cisco Systems Inc. (Nasdaq: CSCO).
The Grating Light Valves developed by Silicon Light Machines bear little resemblance to the MEMS devices used to tilt arrays of mirrors in optical switches. They’re about an inch long and a millimeter or two wide and comprise hundreds of ribbons stretched side-by-side over a shallow trough etched out of a sandwich of metal oxide and silicon.
When no force is applied to the tightly stretched ribbons, they act together like a single, flat mirror. But when alternate ribbons are pulled downwards by electrostatic forces, the two surfaces reflect light at different angles and wavelengths. The light intersects to create an interference pattern or grating, a fundamental building block for all sorts of devices.
As the ribbons are very slender -- their length-to-deflection ratio is around 2000 -- only tiny forces are needed to pull them downwards. Their deflection can be controlled very accurately, which translates into very fine control of wavelengths and gratings, according to Robert W. Corrigan, Silicon Light Machines’ chief operating officer.
The low force and tiny amount of deflection required also translates into very fast switching speeds -- 20 nanoseconds from full on to full off -- and very high reliability, Corrigan says. “We’re barely moving these things,” he notes. “It’s like plucking a guitar string.”
Initially, Silicon Light Machines targeted non-telecom applications for its Grating Light Valves. They're already being used in printers by a couple of manufacturers that together represent 70 to 80 percent of some printer markets, according to Corrigan.
Last summer, Silicon Light Machines also clinched a deal with Sony Corp. that gives Sony the exclusive right to use Grating Light Valves to develop new types of displays. The basic idea is to aim light from a laser at one of these devices incorporating 1080 ribbons -- one for each row of pixels on a giant screen. The device is turned to create the row of pixels at the same time as the ribbons are moved to change the color of the light. The result is an ultra-low-cost 1080x1920 pixel image operating at 60 frames a second.
Sony "is betting hundreds of millions of dollars" on this idea, according to Corrigan, who declines to say how much the contract with Sony is worth. One analyst, who asked not to be identified, puts the value of royalty payments over the next couple of years at $50 million -- not enough to make a big difference to Cypress’s financial prospects.
Cypress acquired Silicon Light Machines soon after the deal with Sony went through. It was already making Grating Light Valves for Silicon Light Machines, and it will continue making them for Sony.
As Sony now has exclusive rights to use Grating Light Valves for most types of display, Cypress has started looking at other applications for the technology, including optical components for telecom gear. Corrigan names very large scale VOAs, tunable filters, optical switches, chromatic dispersion compensators, and PMD (polarization mode dispersion) compensators as possible products but says it’s too early to give specifics on which products will be delivered on what timetable.
Analysts aren’t expecting Cypress to ship any of these possible products until mid 2002, so there’s no immediate prospects of its Grating Light Valves giving the company’s earnings or revenues a boost. Certainly, the stock hasn't benefited. Cypress’s share price is trading close to $20, giving it a valuation of about $2.6 billion. Its price-to-earnings (P/E) ratio is a mere 11.06, according to Stockpoint.
Such a valuation implies that investors have ignored the potential of Cypress's initiatives in optical technology. Pure Optical companies trade at much higher valuations -- sometimes as high as 20 times annual revenues. Consider, for example, that optical components startup New Focus Inc. (Nasdaq: NUFO) is currently trading with a market valuation of $1 billion. For the quarter ending March 31, New Focus booked only $40.8 million and lost $86 million. Now consider that Cypress booked $262.3 million in revenue and generated net income of $10.5 million during the quarter ending March 31, but trades at a valuation of about $2.6 billion.
Cypress’s acquisition of Silicon Light Machines is part of an effort to restore the company's fortunes by ramping up its business in high-margin telecom components and reducing its reliance on low-margin memory and clock chips.
A couple of weeks ago, rumors were circulating that Cypress was considering a second move on this front -- the acquisition of Bookham Technology PLC (Nasdaq: BKHM; London: BHM).
Cypress and Bookham decline to confirm or deny the rumor, but the idea holds water in three respects. First, Bookham and Silicon Light Machines might have synergies because both companies have deliberately picked CMOS (complementary metal oxide semiconductor) process technology to make their widgets so that they can make use of existing fabrication resources to automate manufacturing. Second, Bookham's technology and management is starting to get good reviews from equipment manufacturers, even though its name is still mud in the financial community. Third, Cypress has $1.1 billion cash on its balance sheet and thus is in a position to make an acquisition.
Financial analysts, however, doubt whether an acquistion is in the cards, noting that this isn't the first time that rumors have circulated about a possible takeover of Bookham.
Cypress would have to pay about $1 billion to acquire Bookham, says one analyst following Cypress, who requested anonymity. “At that level -- $1 billion -- any acquisition target would have to be very profitable and would probably have to have just gone public,” he says. Bookham is too big and is struggling to reach profitability, he says. In other words, it's a non-starter.
— Peter Heywood, Founding Editor, Light Reading