To Mr. Gottlieb,
SLIC is no SDLI. SLIC has no real products, only pilots and a lot of hype. They have played the press like an instrument and sucked in some investors. Even their hype doesn't make a lot of sense to me. They boast about their MBE production technique being high yield, but epi yield has little influence over production costs because it's the first step. Throw the wafer away and grow another one. What costs you money is back end of the line yield problems (because you lose all of the work you've invested through that point) and SLIC has no special know how here to boast about. They licensed University work with University lifetesting data from a Professor who is the biggest self promoter since Ross Perot, but has long ago spent her credibility among colleagues. Proper laser lifetesting requires hundreds of devices on test for thousands of hours each. No university can do that. It costs millions of dollars. What a University does is take the best single device they can make and run it. These aren't statistics and no customer's going to put such a laser in the ground knowing they might have to dig it up later amid a network outage.
SDLI has been in business for more than 10 years, and even when they were founded as a spin-off, they took years of know how and world class people with them from Xerox PARC. SDLI has had products for years, many unique in the high power area, and many long standing customer relationships. No newcomer is going to eat their lunch immediately. It will be years and probably a recapitalization before SLIC even registers a two digit PE.
You refer to SDLI's out of date technology, but you should be more specific. If you mean MOVPE growth instead of MBE, 90% of the world uses this technology for optoelectronics because it offers higher throughput and lower cost of ownership. If you mean the AlGaAs instead of the InGaAlP material system, please note that the only two commercial suppliers of near IR power lasers with lifetimes documented over millions of test hours such that they can quote their device reliability in terms of fits (failures per billion hours, typically 100-200) are using AlGaAs.
No major company today manufactures GaAs lasers outside of their research labs because the technology is still too difficult to transfer, even within company. This is true for Lucent, IBM (now Uniphase), NT, HP, Siemens, and is true for all of the large Japanese opto companies. This is my main cause of my skepticism against SLIC. They've taken a motley crew of scientists, some excellent I'm sure, and licensed a university process. I don't expect them to reproducably produce anything state-of-the-art for several years at the earliest. Lasers are extremely complex to design, optimize and produce. I'd rather put my money down in a couple of years once SLIC surprised me and has shown they can do the job.
Yield problems are part of the business, which is a high tech, pushing the envelope kind of business. SDLI will be back on track long before SLIC comes to market with a device they can guarantee has 100,000 hours mean time to failure. I know this because you have to lifetest many devices for 2 or more years to reach that figure with statistics a customer will believe. SLIC hasn't been a company that long. |
