SAN JOSE, Calif.--During the TSMC 2011 Technology Symposium here Tuesday (April 5), Morris Chang, chairman and chief executive of the silicon foundry giant, gave a keynote and discussed several subjects.
Here are seven major points that Chang made during the keynote:
1. Japan impact
The recent quake in Japan will have "no noticeable impact after a lot of work" for TSMC. However, Chang said the quake did impact customers' supply lines and customers' customers' supply lines.
As a result, there will be "some effect in the second quarter, perhaps in the third quarter." The impact will last "primarily one quarter" but no more than "two quarters," Chang said.
2. Semi forecast
The March 11 earthquake in Japan, coupled with other factors, will impact the overall IC business. The IC market, excluding memory, is expected to grow "4 percent" in 2011, down from TSMC's original forecast of "7 percent" just three or four months ago.
The market has seen a "slight softening of the world economy," some inflation and "a few problems in Europe."
The PC and cell phones have been the big drivers for the IC industry. "Now we see a third 'killer app.' It's mobile products. It's smart phones and tablets," Chang said.
4. 28-nm update
TSMC has had 75 tape outs on its 28-nm technology, according to Chang, more than its 40-nm process technology had at a comparable stage.
Regarding TSMC's long-awaited, high-k/metal-gate scheme at 28-nm, the technology is "qualified" and "production ready," Chang said.
5. 3-D chips
TSMC has poured "significant R&D" in 3-D chips using through-silicon vias (TSVs). The company sees it as a paradigm shift called "systems-level scaling," Chang said.
"There are still a lot of challenges for 450-mm," Chang said. Still, TSMC reiterated its position that it would build a 450-mm pilot line in the 2013-2014 time frame, with production due in 2015-2016. The intercept point is "20-nm," Chang said.
7. The competition
Without naming names, Chang said TSMC says is still the leader in process technology among foundries. "Some our competitors have spent a lot of money but they do not have the technology," Chang said.
At the Symposium met a lot of ex Intel types of Chinese origin who are now working for TSMC. Their development methodology seems to be quite similar. We did not get to hear much about the yield / ramp issues except there was a graph that indicated that Defect Reduction might take 6 mo.s or so.
@Warren, right on! TSMC has been mum on the 28nm yeilds! Dan Nenni of SemiWiki was polling some time ago on LinkedIn in a guessing game of TSMC's yield at 28nm, you must have seen that.
@Mark LaPedus: I see no mention of MEMS; it would have been nice to know how many more designs can be transitioned to CMOS MEMS from the 'traditional' MEMS foundries.
Dr. MP Divakar
There might be an eighth point. Dr. Chang said that he projected the industry growth at 5% per year over the next five years. Even though the unit growth would be lots more, 5% is more like a mature industry. Since I can remember periods of 40% growth per year, this sign of maturing is a bit sobering. I guess we could say that TI's acquiring National is another sign of maturing.
According to iSuppli semiconductor revenue will grow by 7.0 percent, up from a 5.8 percent figure previously. But TSMC belives "4 percent" in 2011, down from TSMC's original forecast of "7 percent" just three or four months ago. Contradictory views Not sure who is right ?
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.