As Spansion drives its flash
memory products further into embedded markets such as automotive,
telecommunication, industrial and CE products, “Our expanded deal with
XMC becomes critical for our NOR flash memory products in the future,”
said Jackson Huang, vice president at Spansion. XMC, already proven as a
faithful partner to AMD, has been augmenting products manufactured at
Spansion’s own fab in Austin, Texas. “XMC is important to Spansion’s
fab-lite strategy,” Huang added.
Huang also added that as a graphical
user interface in embedded systems gets more complex requiring fast
response, “we need to be able to offer higher density, higher
performance, cost-effective NOR flash at 32-nm node.”
XMC’s fab is
qualified to produce Spansion’s embedded NOR flash products for the
automotive market, getting all of its certifications in order. The
performance level [of embedded NOR flash] can be maintained at high,
even extreme temperature. XMC also promises the longevity [as long as 10
years or more] of such NOR flash, required for telecommunication
Perhaps most important in the XMC makeover is its
completely new management team. Roughly half the top executives have
international backgrounds. Its CEO Simon Young cut his teeth as R&D
and manufacturing director of logic technology at Intel. He later worked
at Chartered Semiconductor as CTO. Then Young became COO at SMIC.
Mei, responsible for XMC’s technology development, previously worked at
IBM, Philips and NXP, and later served as CTO of Huahong NEC.
the ex-IBMer, helped the Big Blue establish IBM Common Platform and its
partnership strategy. He takes pride in that the playbook he helped
write for IBM is now being applied to XMC – in fostering its partnership
strategy. Lange also served as VP of marketing at Chartered
Semiconductor. Through his Chartered days, he got to know Young, who
gave Lange a call last summer and asked him to join in Wuhan.
XMC has about 1,000 employees including 20 overseas experts. “Compared
to Shanghai, the turnover of our workforce is much lower, and the number
of expats working for XMC is dropping since we can now hire engineers
locally,” said Lange.
If they are doing straight cmos from 0.35 to 65...i see this venture to be a gross failure...unless they invest R and D and look at doing non standard stuff line SiGe BiCMOS or GaN...this venture will get eaten by the likes of TSMC, GloFo....
They know it. That's exactly why Lange, ex-IBMer and now at XMC, is saying that's not what the new foundry will do. Instead, XMC is focused on selected customers, forge "partnerships" (a la IBM Common Platform), customize process, and develop "technology with a twist" for that partner, as he put it.
"Megafoundries" are only needed to service a few, monster markets like smartphones and tablets. There you have to scale and bleeding edge technology to make the cheapest memories or processors to serve huge volume orders. It is a total commodity play. There's a lot more innovation in the semiconductor industry than just pushing Moore's Law and I think XMC and others realize this.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.