Digitimes reports that the proportion of Samsung image sensors used in HTC handsets is expected to reach 50% with a more competitive pricing at the end of 2009 amid sales of HTC's 5MP smartphones. Currently 60-70% of HTC's handsets use Aptina's products, according to Digitimes' sources.
The funny thing is that Samsung is Korean and HTC is based in Taiwan. So much for the country-based supply chain theory. If that were true, HTC would be going with Omnivision.
ReplyDeleteThis is clearly another bad data point for Aptina. I'm surprised Micron hasn't shut it down yet. Do they still think they can sell it?
For me, having Nic on-board means the chances to sell are still there. Otherwise, Nic would leave.
ReplyDeleteWhere's Nic gonna go? Not too much work for vulture capitalists these days. As long as he can continue to convince Micron's BOD that Aptina is sellable, he keeps his job.
ReplyDeleteHow much longer can Micron continue to spend $30mm per quarter on imaging? Aptina is no longer a source of cash flow, that's for sure.
Well, I guess the current Aptina's price might not please Micron's BOD, but it's certainly not zero. So, Nic only should convince BOD that the price would rise in future - an easier task.
ReplyDeleteRegarding Aptina's losses, I wonder what part of them is caused by the way Aptina pays to Micron for production. I suspect that Aptina is paying for fab capacity, rather than per wafer price. This would explain last quarter results, where production went down by half, while margins were down by a factor of 10. If my suspicion is correct, shutting down Aptina would just move its loss to Micron books. One needs to sell Italian fab to get rid of loss - not easy these days.
I think Micron would just shut down the fab, but that's probably expensive. You'd have a non cash charge for whatever upgrades they did to the fab to get it to run image sensors, and then they'd have a cash charge for the laid off employees. Not sure what the labor laws are like in Italy, but in most of the EU downsizing is not easy.
ReplyDeleteOTOH, I'm not sure that Nic can make much of a case for things improving. If these rumors are right, things are getting worse at Aptina, not better.
Don't get me wrong, I feel really sorry for the folks over at Aptina, but the industry does need to consolidate and it will consolidate.
consolidate with ST?? Is it true??
ReplyDelete1. OV cannot provide good sensor to HTC
ReplyDelete2. Samsung can get in because if its panel
3. Samsung indeed gets some 5M sockets in HTC, however, Samsung's 5M still has some problems...
Digitimes, always has its preferred bias...
Could you explain your statement #2? It seems incomplete.
ReplyDeleteSamsung provides LCD panel to HTC. Thus, Samsung usually tries to have the bundle sales at HTC
ReplyDeleteAptina + STM imaging? LOL. Hint: Adding two high cost manufacturers together does not make them more efficient. Would STM even be in this business if Nokia didn't insist?
ReplyDeleteAccording to your logic, Toshiba and Sony should be out of business too, as manufacturing is very expensive in Japan.
ReplyDeleteThere were reports that Aptina and ST are looking for second source fabs on the Far East. Other reports say that Aptina has settled on Nanya as its second source. If this is correct, Aptina might soon have very cheap manufacturing.
I don't know if adding high cost manufacturers together make them more efficient but I know that some "low-cost" manufacturers deliver crappy images and technology bullshits. Some might like it, some may not.
ReplyDelete"but I know that some "low-cost" manufacturers deliver crappy images and technology bullshits"
ReplyDeleteIs that a back-sided swipe at the great Omniscience?
I think many people underestimate how much the fab process contributes to image quality. It would be very hard for Micron/Aptina/ST or anyone else for that matter to move their sensors to a different fab and still obtain the same quality at short notice (I'd say a good tech transfer would be 2-3 years). There are only a handful of fabs in the world with the right technology that can produce reasonable quality CMOS sensors below 2.8um pixel sizes.
ReplyDeleteOff the top of my head, the only proven fabs are: TSMC (Taiwan), Toshiba (Japan), Sony (Japan), ST (France), Aptina (US/Italy), Samsung (South Korea), Dongbu (South Korea)
There are of course other fabs, but none of them can compete in ultra large volume prodution as the ones I've listed above and as such they remain niche players and are usually behind technologically.
Yes, technology transfer is difficult. Even equally equipped fabs within the same company running supposedly same process often differ in image quality. And even a "good" fab is unable to sustain the same image quality when its utilization drops, as many of us experience these days. And some say that even different shifts at the same fab differ in yield and image quality.
ReplyDeleteIndeed, history knows unsuccessful attempts to transfer CIS process, despite many years spent. However, there were some success stories too.
All in all, I agree with you that a successful process transfer takes many years. Still, probably after one year or so the recipient fab can start producing cheap low-quality VGAs and 1.3MPs for emerging markets and gradually move to higher resolution higher quality products.
and 2+ years later, Aptina is still going strong! The naysayers can shove it!
ReplyDelete