Silicon Forest silicon manufacturing: Still the world’s best

Among many interesting developments announced at the Intel Developer Forum (in San Francisco, where the technology discussion was largely about what Intel is doing in Oregon) was the announcement that Intel’s somewhat understated Custom Foundry business would be building processors for ARM and its customers, including LG, servicing the mobile and IoT markets. While this is not as big as a win as having the dominant branded design, it underscores Intel’s continuing leadership in manufacturing technology—led by Intel’s Hillsboro Ronler Acres site.

And what leadership it is! Industry insiders know that semiconductor technology “nodes” are aggressively labeled (i.e. actual lines and spaces in the current 14nm node are closer to 25nm) but also that Intel’s nodes are higher in density and performance than those of competing manufacturers (such as Samsung, TSMC, Global Foundries). What Intel calls “10nm” is closer to what others call “7nm”. Intel also announced that it would be extending its technology performance lead (from 12-18 months or so to 18-24 months) at the 10nm node in 2017 and that classical Moore’s Law expectations of decreasing cost per transistor were still in force, at least at Intel, if not for the entire industry.

Cost per wafer is certainly going up as expected, but density enabled by Intel’s FinFET device technology (some detail behind link above) is increasing much faster. Hence Intel’s strong lead in processors for data centers – and the need for other companies (including mobile processor competitors) needing the absolute best technology to get it from Intel, and that first from Oregon. Intel’s Custom Foundry is both an expression and extension of its technology leadership, where keeping perfectly-running $20B fabs full is crucial for company financial performance.

A major driver of increasing wafer cost is lithography –the price and productivity of major tools, and the number of times (mask layers) they are used to complete a wafer. One of the favorite parlor games in the industry is when EUV (Extreme Ultraviolet Lithography) will come on line. Intel remains coy about this, keeping pressure on equipment and materials vendors. In the meantime, Oregon startup Inpria has announced that it is scaling up production of its EUV photoresist, using the proceeds of its recent $10M+ investment round. All signs currently seem to say: EUV will happen (including the possibility of retrofit) and the timing will be based on yield and uptime, i.e. cost.

High performance electronics in attractive small packages (like phones) requires packing lots of different chips very tightly—including in multi-chip packages. Intel’s Custom Foundry also offers an elegant entry in “2.5D” integration called EMIB (Embedded Multi-die Interconnect Bridge) where smaller silicon interposers are embedded.

Meanwhile, Oregon startup Nano3D Systems is making great progress on novel, inexpensive plating technology to enable true 3D packaging and more. They report that they are now qualified as a supplier to a “major” global chemical company and also the world’s largest distributor of chemical materials. They will report progress on their general/copper and INVAR plating technologies at the fall Electrochemical Society Meeting. Like EUV, this may sound esoteric (especially for a conference in Hawaii) but it is actually breakthrough stuff for electronics miniaturization and performance.

The bottom line here is that semiconductor and electronics manufacturing remains a hot and dynamic field, and nowhere more so than Oregon, where major elements of its future are being invented.

If you read this far and still wish for more…

To provide an idea of how important all this is to Oregon, I recently completed a “state of the industry” report for Business Oregon, which Silicon Florist has posted below. The report contains economic statistics and a bit more technical depth on trends in the industry.

ONAMI: Computers and Electronics State of the Industry by Rick Turoczy on Scribd


Skip Rung is a senior high-technology research and development (R&D) executive with over 25 years of engineering and management experience in complementary metal-oxide-semiconductor (CMOS) process technology, integrated circuit design and packaging, MEMS, microfluidics, and inkjet printing. Skip currently serves as President and Executive Director of ONAMI.