Tech executives, researchers and government officials are gathering in Seattle this week to figure out ways to add a new dimension to America’s chip industry — figuratively and literally.
“We’re going to talk about a once-in-a-lifetime opportunity to reinvent domestic microelectronics manufacturing,” Mark Rosker, director of the Defense Advanced Research Projects Agency’s Microsystems Technology Office, said today at the opening session of the ERI 2.0 Summit at the Hyatt Regency Seattle.
More than 1,300 attendees signed up for the DARPA event, which follows up on a series of Electronics Resurgence Initiative Summits that were conducted before the COVID-19 pandemic.
The main objective of this week’s summit is to work on ways to boost research, development and manufacturing for the chip industry. DARPA is just one of the government agencies involved in such efforts. Representatives of the Commerce Department, the U.S. Department of Energy, the White House Office of Science and Technology Policy and the National Science Foundation are also at this week’s summit.
DARPA’s role in supporting innovations in microelectronics is funded through the Department of Defense, but other parts of the federal government are drawing upon $52 billion in funding provided by last year’s bipartisan CHIPS and Science Act. Sen. Maria Cantwell, D-Wash., told attendees that getting that legislation through Congress wasn’t a trivial task.
“I can tell you it really was a herculean effort, because it’s a lot of money,” she said. “And not only is it a lot of money. We had to explain to people what happened in the last two decades, that all of a sudden now we’re saying we need to bring the supply chain [for microelectronics] back to the United States of America.”
Cantwell said the pandemic and its effects revealed how much needed to be done — a factor that Intel CEO Patrick Gelsinger brought up as well.
“COVID was a massive wakeup call,” Gelsinger said. “Imagine we’re short a $2 semiconductor and we can’t ship a $30,000 car or a $15 million jet, right? All of a sudden we realized how fragile our supply chains had become.”
As of last year, only 12% of the world’s chips were made in the U.S., which is a steep drop from the 37% figure for the 1990s. Today, about 80% of today’s chips are made in Asia — which left America vulnerable when the pandemic dealt a heavy blow to trans-Pacific supply chains.
The federal government’s chip initiative aims to change that equation, primarily by providing incentives for chipmakers to build new manufacturing facilities in the U.S. That strategy seems to be working: Gelsinger pointed to Intel’s multibillion-dollar investment in new chip factories built in Arizona, New Mexico, Ohio and Oregon.
Innovations in chip design are also playing a part. “We are finally at the end of the age of the monolithic integrated circuit,” Rosker declared. That’s not only because chipmakers are approaching the physical limit for cramming more transistors onto a tiny wafer, but also because of the economics of the industry, he said.
Rosker said a DARPA initiative called Next-Generation Microsystems and Manufacturing, or NGMM, offers a path for making the transition from 2-D to 3-D microelectronics.
“We imagine layers of electronics with super-dense interconnects that provide paths for information to travel just as easily vertically as they do today horizontally,” he said. If the initiative turns out the way DARPA hopes, that would lead not only to faster chips, but also to new types of materials and devices.
The industry is already shifting from mass-producing monolithic chips to creating “chiplets” that can be packaged together in three dimensions like Lego blocks for customized applications.
“Essentially, old silicon stuff is becoming cool new packaging stuff,” Gelsinger said. “We’re now entering into the advanced-packaging era, where 2-and-a-half and 3-D packages become the new norm.”
And it doesn’t have to be just the old silicon stuff: Rosker said DARPA is “looking at ways to incorporate photonics and non-silicon electronics into these systems.”
DARPA is interested in giving the semiconductor industry a boost because the U.S. military is one of the industry’s most demanding customers. The Pentagon wants to make sure that it has a reliable supply of chips, that the hardware won’t be vulnerable to foreign hackers, and that the chips can stand up to extreme conditions such as space radiation or the high temperatures in jet turbines.
Federal efforts to support the chip industry are just getting started. Some programs, such as the National Semiconductor Technology Center and the Pentagon’s RAMP-C program, are still ramping up.
Cantwell, meanwhile, is already thinking about future initiatives. She called for DARPA-funded researchers to be encouraged to join the NSTC, and for more partnerships to be forged between DARPA and semiconductor industry innovators.
She also said the executive branch should develop a government-wide workforce strategy for the chip industry. Cantwell noted that NSF recently awarded $10 million to the University of Washington to support semiconductor workforce development and research — a project funded by the CHIPS and Science Act.
“We did two competitiveness bills before this legislation, and we failed to do one thing after it: appropriate the right amount of money,” Cantwell said. “So I hope, out of today’s conference, that you will also focus on the fact that Congress needs to stay competitive and continue the resources besides the $52 billion for the semiconductor grant program — but making sure that we get the scientists for tomorrow.”
Correction for 4:40 p.m. PT Aug. 23: This report has been revised to clarify the roles being played by DARPA and other federal agencies in supporting innovations in microelectronics — and to make clear that DARPA receives no funding from the CHIPS and Science Act.
