The U.S. government should launch a new multi-billion national research effort to define a post-silicon semiconductor industry.
The working group of semiconductor veterans the White House recently convened should recommend the initiative in its report expected early next year. It could give the next U.S. president a rare opportunity to kickstart an engine for a new golden age in high tech.
When it launched the working group, the Obama administration cited challenges the U.S. chip sector faces with the slowing pace of Moore’s law and rising competition from China. The country is bankrolling its own industry with a multi-billion dollar fund.
While China promises to be a real threat to U.S. chip companies, the looming limits of today’s semiconductor technology are a much larger common enemy. The U.S. has an opportunity to pioneer a path beyond the 3-5nm process technologies companies see at the distant edge of their road maps.
A handful of veteran industry consultants and market watchers agree, lending numbers to the argument.
The U.S. should create “a national lab like Imec in Flanders or CEA-Leti in France but much bigger, driving U.S. tech leadership—maybe the mission is to develop a 2nm node,” said Handel Jones, chief executive of International Business Strategies (IBS).
The lab should be founded with an initial outlay “north of $10-20 billion” and an annual budget of $2-4 billion, Jones said. “We have to think big,” he said, noting the funds would pay for the capability to prove new technologies can be put into production. “I think it’s important we have high volume manufacturing of advanced components,” he said.
A national fab also could act as a so-called trusted foundry to make limited runs of ASICs for military projects. IBM’s fabs used to play this role, but now they are part of the foreign-owned Globalfoundries.
The presence of Wes Bush, chief executive of Northrup Grumman, on the White House panel is a nod to the need of a trusted foundry, said Risto Puhakka, president of market watcher VLSI Research. Northrup Grumman has been rumored to be considering building such a fab for government projects.
The government’s needs for custom chips “are probably not very big and not very advanced,” said Puhakka, making the trusted foundry a small but strategic part of the plan.
Overall a national lab “something like the former Sematech effort in targeted semiconductor technologies…would be a good thing,” said Rob Lineback, senior analyst with IC Insights.
“When Sematech was formed in the 1980s, the Japanese were seen as the major threat to the U.S. chip industry,” Lineback said.
Today China forms an even more daunting competitor in terms of the size of its wallet and market. However, “I think the big challenge is to find common ground for U.S. semiconductor companies--it won't be easy,” Lineback said.
Next page: R&D declines as the landscape broadens