When Intel (INTC +7.06%) announced that it would be partnering with Rockchip to build low-cost tablet processors, I believed that the deal was a "smart, bold move." Then, when rumors surfaced that Intel would be taking an equity stake in Tsinghua Unigroup, which owns Spreadtrum and RDA Microelectronics, I noted that Intel could be doing this deal for similar reasons, with the same ultimate strategic goal: proliferate Intel's X86 architecture and, in due time, fill Intel's factories with chips built under this partnership.
However, Bernstein Research's Stacy Rasgon notes that his firm's Asian team "believes potential pressure to use SMIC (SMI +0.00%) in the longer term may serve to blunt some of Intel's manufacturing advantage."
Let's explore this in more depth.
Who is SMIC?
China-based Semiconductor Manufacturing International, or SMIC, is one of the world's leading contract chip manufacturers. According to IC Insights, SMIC was the fifth largest chip foundry during 2013 by revenue. It also saw revenue growth of 28% during 2013, making it one of the fastest growing contract chip builders.
However, one thing about SMIC is that it's not really a leading-edge foundry. For example, SMIC is now just transitioning to 28-nanometer chip manufacturing technology, which companies such as Taiwan Semiconductor (TSM +1.32%) and Samsung have been producing for quite some time. This means that for products that don't need the latest and greatest, SMIC is viable.
However, for bleeding-edge, highly competitive products, SMIC probably isn't a terribly good choice.
What does this have to do with Intel and Spreadtrum?
My initial hypothesis was that Intel would try to get Intel-designed processor cores into as many Spreadtrum chips as possible, and then eventually Intel and Spreadtrum would transition the manufacturing of future designs to Intel-owned manufacturing technologies and factories.
However, as Rasgon points out, Spreadtrum may be pushed to use SMIC for a number of applications since Spreadtrum is owned by Tsinghua Unigroup, which is a company owned by the Chinese government. Given that the Chinese government has publicly signaled that it wants to support China-based chip designers and manufacturers, Rasgon and his team certainly have a valid point.
One could conceivably argue that Intel could eventually transition these chips to its China-based Fab 68, but given that this factory is still being used to build 65-nanometer technology (that's four-generation-old tech), it's not likely that Intel is going to outfit its China-based factory with its latest and greatest.
So Intel and Spreadtrum may ultimately end up building many, if not all, of the future chips from their partnership at SMIC.
Why this is still a strategic victory for Intel
Even if Intel doesn't use these partnerships to drive incremental volume into its factories, it's still a strategic victory. Let me explain why.
Intel is fundamentally fighting two separate but related battles. First, it's fighting an architecture war against ARM Holdings (ARMH +0.00%). As long as ARM has dominant market share in the smartphone and tablet market, developers are going to prioritize ARM-based Android applications first and then will give Intel Architecture a look if they're so inclined. Having tons of Intel Architecture-based chips out there means that the developers will have no choice but to optimize for Intel, too.
Next, Intel is fighting an economies-of-scale war with the foundries. The more leading-edge volume that it can drive into its factories, the less competing foundries will be able to afford to invest in future bleeding-edge technology.
If Intel can move a ton of smartphone volume to both Intel Architecture and to its own factories, then this keeps ARM and the foundries at bay. If Intel doesn't build these chips in-house, then it still has the benefit of dealing a blow to ARM Holdings.
Foolish bottom line
I have no clue what Intel's plans are going for the manufacture of these chips. My guess is that based on the "second half of 2015" availability for the chips from this partnership, these chips will probably be built at Taiwan Semiconductor, which is where Intel CEO Brian Krzanich has indicated that the company's in-house SoFIA (and chips built as part of the Intel/Rockchip partnership) will be built.
Where future chips ultimately end up being built is still an unknown, but whether they're built at Intel or elsewhere, Intel still at least gets one in on ARM Holdings.
