|
Intel Exits LCOS Business -
The Real Story
|
| Intel (Santa Clara, CA) (www.intel.com)
has finally confirmed the rumors we have been hearing for the last
month or so - the company will officially close down its LCOS development
efforts. After reviewing the project, Intel management decided that
the opportunity for LCOS microdisplays was not large enough to justify
the continued investment necessary to fully commercialize the technology.
Perhaps two-dozen jobs will be affected in California, Oregon, New
Hampshire and Arizona, with most being offered new assignments. The move follows the company's decision to refocus effort from a 1280 x 720 panel to a 1920 x 1080 resolution panel last August (http://www.insightmedia.info/news/IntelShifts.htm). It also follows recent high profile problems with the Sears LCOS-RPTV launch (http://www.insightmedia.info/news/SearsOffersVeos.htm) and the shut down just two weeks ago of Philips' LCOS operations (http://www.insightmedia.info/news/PhilipsClosesLCOSOp.htm). Is this the end of the road for LCOS? We think definitely not. All LCOS approaches are not created equal and the termination of Philips and Intel's LCOS effort is more of an indicator of failed methodology than a failed technology. Sony, JVC and a number of other LCOS players, both public and in stealth mode, have better technical methods that should yield successful products. However, the fact that Philips and Intel have exited the LCOS arena will cause many to conclude that LCOS cannot be commercialized. This could make it tougher for the remaining LCOS players to gain confidence in the industry and in the minds of consumers. So what went wrong at Intel? We see three critical factors; problems with the LCOS panel design, a shifting market for LCOS panels and a changing business model for LCOS commercialization within Intel. All of these factors skewed the opportunity index for commercializing LCOS leading to the business decision that the cost of continued investment would not be outweighed by the revenue potential. Let's start with the LCOS panel design. Those who had evaluated Intel's LCOS design found it a brute force type of design with way too much circuitry. In fact, this circuitry was so massive that it extended out beyond the active area of the display. This means a large die, fewer displays per wafer, higher costs and lower yields. Intel chose this approach to include memory on the backplane to simplify the interface, but there was a trade off. This design may have been the result of how the company staffed the LCOS project. According to some sources, about 75 personnel were assigned to the LCOS program via Intel's internal redeployment program. Perhaps 25 came from the semiconductor side; maybe 20 were assigned to solving the fundamental LCOS development issues, with rest assigned to understanding system design and integration issues. As a result, the backplane design was heavily influenced by the silicon team yielding an over-sized, far too expensive approach. As we reported last August, Intel initially planed to offer 1280 x 720 resolution LCOS panels for RPTVs, but then changed its focus to work on first commercializing a 1920 x 1080 resolution panel. This change reflected the market reality that most customers already had a good 720p solution and wanted LCOS to solve the 1080p problem. This strategic move meant more investment and a shift in expectations for revenue generation. In addition, the shift tied Intel to the success of 1080p products in the market. When this will occur remains an area of some speculation. As a data point on this, we asked the attendees at the HDTV Forum last August to tell us, "In what year will sales of 1080p RPTVs equal those of 720p RPTV?" Of the 160 respondents, 44% thought that year was 2008, with 16% thinking 2006 and 18% thinking 2010. If accurate, that means by moving to a 1080p approach, the majority of the revenue for Intel would have shifted out two years. That's enough to kill any business model. Finally, the company's internal business plans to commercialize LCOS changed along the way. Apparently, when Intel first began to develop LCOS, the idea was to sell panels and light engines to potential RPTV system integrators. At some point, plans changed to only sell the LCOS panels, which meant higher volumes would be needed to justify the investment. Gross margins on most of Intel's silicon products are expected to be 50% to 60%, but Intel needed to offer panels at competitive pricing in a market with fast-declining prices. Add to this environment the pressure of developing and commercializing a new technology - and doing it within a defined window of opportunity. Intel was actually doing a good job working with partners to develop 1-, 2- and 3-panel engine designs as well as multi-primary color systems. Some of these prototypes had shown very good performance levels. Plus, history has shown that a full engine reference design is absolutely essential to commercializing a display technology. All of this effort was put on hold following the August 1080p shift. Management finally added it all up and decided the revenue potential did not outweigh the risk and investments needed to be successful. It decided resources were better spent in other areas of greater potential return - a similar conclusion that Philips' management came to. We estimate Intel probably spent north of $50M on LCOS over the last 5-6 years. Plus, it spent tens of millions more in related projection technology and company investments. Some of these investments remain, but Intel is still in the process of deciding if and what it will do with the IP and know how it acquired. This latest move just adds to the bad year that Intel in having. In addition to canceling the LCOS program, the company canceled the 4 GHz Pentium 4 chip that was supposed to ship early next year and suspended work on a next-generation Pentium 4 design. It faced a small recall of some chips and then delayed the release of Alviso, a mobile computer chip. Even when thing appear to go well, such as higher yields on a Pentium chip, the inventory built up too fast hurting business. Go figure. Given all of the execution issues, there is a clear need to focus on the core products at Intel. LCOS just did not offer enough bang for the buck to make it above the cut-off line. |
|
|
To be removed from the Insight Media News list, please reply with UNSUBSCRIBE. |
Copyright©
2004 Insight Media
Offers and prices subject to change without notice.