"CIS is an analogic variant of the CMOS process commonly used for memory and logic circuits, and has become a key segment in semiconductor, reaching $15.5B in 2018 and exceeding 3% of total semiconductor sales.
In 2019, the overall attachment rate for CIS cameras per phone is moving towards in average of 2.5 units per phone, and the growth rate for CIS attachment will rise from 6.5% to 7.8% from 2019 – 2021. Amidst stagnant smartphone volume, CIS attachment rate is a central, successful strategy for main smartphone OEMs like Apple, Huawei, and Samsung. As camera quantity and die size increase per end-device, a 10.1% year-on-year growth rate is expected for 2019.
Alongside mobile, which is the main application market (representing 70% of all CIS sales), security and automotive are experiencing double-digit growth and have grown into billion-dollar CIS segments. Again, the attachment rate per endsystem is the key metric to monitor.
2019 looks slightly different than 2018. With a low Q1, the CIS market faces a slowly eroding ASP since most players can now match Sony’s proposition. Nevertheless, the market remains constrained in terms of capacity, with capex the main limiting factor since customers always want more CIS cameras. The outlook though remains very positive – in the range of 10% YoY in 2019 and 8% over the long-term – CIS is heading for $24B in 2024.
New increases in resolution (16Mp and beyond 20Mp) are linked to new progress in pixel size, but momentum is slow in the 0.8um pixel size range. The consequence is increased die size and silicon. CIS wafer volumes are approaching 250kWpm and will climb to 350kWpm before 2024, necessitating more CIS manufacturing lines to be created or converted from regular CMOS lines."
Seems relevant information at high level... however, how come one can predict a “sudden” YoY drop of grow form 9% to 0% when projecting future changes that will happen from 2023 to 2024... magical crystal ball? This prediction looks rather fishy and not built on any solid base assumptions nor industry trend... this little ‘market point artifact’ is killing somehow the consistency of the full reasoning and data presented...
ReplyDeleteWe should less pay attention to analysis that indeed is supposed to be based on scientific analytics while the real life model is a lot actually closer to chaos theory with high sensitivity to weak parameters. Analyst should not follow journalist/media trends, where buzz makes business and very few of disruptive trends are forecasted clearly enough to be useful.
ReplyDeleteFor sure not any task