There are some trends in the semiconductor world that are heading for an enormous slow motion collision.
From my previous post we know that TSMC is working hard to increase the performance of ARM cores. Presumably the compute performance of these cores is marginal or power reduction would still be the primary goal of the largest ARM core foundry in the world.
On the other hand, Intel’s compute performance is abundant and they are diligently working on power reduction and SoC integration through advanced manufacturing technology.
So, ARM core power will be increasing while Intel core power will be decreasing.
The first touch point for this collision of these opposing trends will come into view in the next few months as Intel ramps up 22nm TriGate and the world sees exactly what the power and performance specs really are. The collision and damage should be complete by the time Intel ramps 14nm TriGate.
Intel is adding 10s of billions of dollars of fab capacity right now. Does anyone out there really think that the general market for semiconductors will increase by $30 billion during the next couple of years and that Intel will get all the increase? Does even the most ardent Intel fan believe that the CPU business will grow by $30 billion anytime soon? No, this is the amount of business that Intel will take from other players in the industry. The big victims will be any fabless semi guy who is dependent on TSMC’s most aggressive technology. The QCOM, NVDA, Samsung, and TIs of the world will be hard hit. TSMC itself will be decimated in 65nm and below. The Samsung facility in Austin will lose their Apple business and be forced to deploy a foundry business model….or sell the facility to a consortium of fabless semiconductor companies. ARMH? They get hurt of course, but then, who cares?
There are big changes coming in the world of semiconductors and most don’t see it coming.
Intel shares will go to $80-100 in this environment. Pretty soon the big money Wall Street guys will figure this out.