SA Author Ashraf Eassa recently wrote an article titled "Intel's real competitor isn't Qualcomm, it's Samsung". He could not be more correct, but not just for the reasons he set out, in my opinion. I see the real war dependent on both process advantage and foundry capacity and only Samsung (OTC:SSNLF) is able to meet Intel (NASDAQ:INTC) on these parameters.
The importance of having foundries in the semiconductor business is ignored by investors betting on sustained success by the "fab less" suppliers like Qualcomm (NASDAQ:QCOM) and others typically using technology developed by ARM Holdings (NASDAQ:ARMH). As process dimensions shrink and chips become smaller and more powerful, capacity to produce the newer and more powerful devices becomes scarcer.
McKinsey did a review in 2009 that pointed out just how little of the worldwide semiconductor foundry capacity was available for process nodes below 32nm.
There is a race going on in mobile computing kicked off by Apple (NASDAQ:AAPL) with its A7 64-bit processor which is now focusing users on computing power rather than just battery life. "Good enough" computing is waning as demand for faster and more powerful chips rises, as pointed out by SA author Terrier Investing titled "Intel: The Inconvenient Truth About Good Enough Computing".
Qualcomm and Samsung have both indicated they will introduce 64-bit processors this year. It is not the 64-bit feature that is driving this train, it is the consumer now believing that "faster" and "more powerful" processors are both necessary and desirable, as was the case in the late 1990s as the personal computer industry vied for customers on the basis of whose PC had the most powerful processor, came with the most random access memory, had the fastest graphics card and so on.
A competition for processing power is uniquely Intel territory. Few competitors can keep in that game and even fewer have the horsepower to compete with Intel.
In its 2009 review, McKinsey also pointed out just how few device manufacturers were able to sustain the investment needed to produce leading edge nodes, concluding that only Intel and Samsung had the resources to invest beyond 22nm.
Intel is already producing 14nm silicon and will have 10nm and plans to reach 7nm and later 5nm in the next few years.
The landscape of semiconductor manufacturing has changed since McKinsey's article in 2009 and continues to change, but what has not changed is Intel's foresight in pushing its process nodes ahead while others based their strategies on "good enough" computing with the view that long battery life and reasonable performance was going to be enough for most users. It was enough and gave smartphone and tablet manufacturers the ability to leverage ARM technology to create a massive industry with worldwide significance and propel Apple and Samsung to the leading positions in mobile.
It will not be enough for very long. When the average user starts bragging about the 64-bit A7 in his iPhone being the fastest and most powerful in mobile, they are creating a customer ethos that once again says power is paramount and that ethos is gathering steam. Just read the comments on SA articles about Apple to get a good cross section of users views on the importance of power now that Apple has made it a feature worth noting.
What goes around comes around. Like Brer Rabbit telling the Fox to "just throw me in the Briar Patch" it seems wittingly or unwittingly Intel now has a playing field where processing power is its "Briar Patch" and suppliers of ARM based processors have entered that space to compete. Intel is well suited to that arena and despite their successes of the past few years neither Qualcomm nor Nvidia (NASDAQ:NVDA) nor Mediatek nor AMD (NYSE:AMD) is equipped for a head to head competition with Intel based on computing power.
For the record, and Intel Core i7 processor today is somewhere between 5 and 10 times more powerful than an Apple A7 depending on what benchmark you use. With a 14nm process and Intel's voltage regulation technology, a processor that powerful can readily fit into a handheld. Intel's early forays into the lower priced mobile market were based on degrading the performance of its chips to sell them at lower price points in Netbooks and tablets, not on some inability to produce a sufficiently powerful processor.
More importantly, even with competitive designs the scarce resource may soon be Foundry capacity. It takes time to build and it's not cheap. Competitors relying on merchant foundries may find the price of playing poker has risen and the market may not be willing to wait for new merchant foundries to be built as it tries to satisfy an insatiable demand for more computing power in handheld devices.
Samsung will continue to have the strategic advantage of full integration but its willingness to act as a merchant supplier to others may be reduced if demand for its own products stretches its ability to produce.
The demand for processors is by no means limited to mobile. Base demand for personal computers seems to have bottomed, server demand remains high and the nascent Internet of Things can quickly become a huge vacuum sucking up all available foundry capacity not already committed if it gathers speed as quickly as it might.
I am long calls on 120,000 Intel shares hedged by a 10,000 share short position.