Intel Announces Major Manufacturing Expansion
Intel has been on something of a tear of late. After its Architecture Day earlier this month, where the company debuted its plans for new CPUs, GPUs, 3D interconnects, and overall foundry strategy, it’s now announcing a new set of manufacturing expansions intended to serve growing product markets.
According to Ann B. Kelleher, senior vice president and general manager of manufacturing and operations at Intel Corporation, Intel will expand its manufacturing capabilities in pursuit of a total addressable market for silicon it estimates at $300B. This shift is being pitched as part of Intel’s transformation from a PC-centric company to a data-centric company. While such terms might seem like little more than marketing, there’s real cash behind the shift. The combined impact of AI, machine learning, IoT, IIoT (industrial Internet of Things), self-driving vehicles, and 5G connectivity make the conventional PC industry look tiny. Long term, these markets could even outstrip whatever slice of mobile revenue that Intel might have claimed a half-decade back. Having missed in that instance, the company is determined not to miss again.
In addition to the 14nm capacity shifts the company has already undertaken this year, Intel will continue its scale-out work in Fab 42, in preparation for 7nm production at that facility. It plans to perform future Optane development at its Rio Rancho facility in New Mexico, along with unspecified plans for next-generation memory and storage technologies. While Kelleher doesn’t mention Optane specifically, the Albuquerque Journal reported in September that Intel was hiring for Optane production at the Rio Rancho facility following its split with Micron.
In addition to this, Intel is now in the planning phase of expected expansions in Oregon, Ireland, and Israel, with construction expected to begin in 2019. Expanding its existing fab space could easily be cheaper than planning to build a new facility from scratch; fab plants aren’t cheap to build and they don’t build quickly, either. But interestingly, Intel isn’t planning to stop collaborating with other foundries, either. Kelleher writes:
In addition to expanding Intel’s own manufacturing capability, we will continue our selective use of foundries for certain technologies where it makes sense for the business. The use of foundries has been an Intel practice for nearly two decades. As we invent more products for a broader set of customers, you can expect us to be strategic about the application of Intel’s differentiated manufacturing capability and the selective use of foundries.
This could be a sign of a long-term strategy shift. It’s true, Intel has used client foundries from time to time over the years. If it simply taps these relationships for manufacturing relatively small amounts of hardware, equipment that’s adjacent to its primary product lines (like chipsets), or for silicon it acquired that was previously built at a different foundry, it wouldn’t be much of a change from what Intel does already. If, on the other hand, Intel were to decide to shift some of its lower-margin manufacturing to client foundries to improve its own margins and did so on an ongoing basis, it would represent a shift from the company’s typical manufacturing strategy. Kelleher seems to imply that Intel might be contemplating this kind of change. If it does, we’d definitely expect Santa Clara to keep its high-end parts and processes for itself, farming out work on parts whose margins are less important to the corporate bottom line.