LONDON – Intel and Samsung will spend $25 billion on increasing their manufacturing capacity in 2013 as this sector of the industry continues to consolidate round very few leading-edge manufacturers, according to IC Insights.
Five companies that are expected to spend at least $3.0 billion in 2013, the same as in 2012 and 2011 and the top-10 capital spenders in 2013 are forecast to increase their spending by 5 percent as compared to 2012, while non-top-10 companies are expected to cut spending by 8 percent.
Over the four-year period 2010 to 2013 Samsung is forecast to spend $46.9 billion, with about 60 percent on its memory production and 40 percent on its break into logic and foundry services. Over the same period Intel is forecast to make $40.0 billion in capital expenditure. "Notably, the combined spending by Samsung and Intel represented 40 percent of the world's semiconductor capital outlays in 2012, with this percentage expected to rise to 42 percent of total capital spending in 2013," IC Insights observed.
IC Insights has also broken the forecast data down by geography. This reveals that, thanks to Intel and Samsung, North America and Korea are of growing significance in chip manufacturing. Japan with 7 percent and Europe with 2 percent of expected capex are more or less out of consideration. Europe's very low figure reflects the almost complete adoption of a fab-lite strategy across the continent.
Taiwan's flat share is due to the second-tier DRAM producers Nanya, Powerchip and ProMOS keeping capex to a minimum while foundries TSMC and UMC strive to be aggressive, IC Insights said.
Click on image to enlarge.
Top ten spenders on semiconductor manufacturing by forecast capex for 2013. Source: IC Insights.
Click on image to enlarge.
2013 share of semiconductor capex by region. Source: IC Insights.
Call me crazy, but I am bearish on Samsung long term. They are way too dependent on internal demand to drive all their capex. They are scaring off their foundry customers by competing with them. The immense profits they are generating in phones can't last forever, the barrier to entry for generic Android smart phones is small, it will attract competitors until their margins drop. Their is a chance of a "house of cards" situation if there is a decline in internal demand.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.