IFTLE 463: DoD Focuses on “Reshoring” Electronics to the US
October 5, 2020
The Defense Advanced Research Projects Agency (DARPA) launched its Electronic Resurgence Initiative (ERI) in 2017 with the intention of reshoring a domestic chip industry that has been moving steadily offshore for decades. Microelectronics are a foundational building block of most of our defense systems
At this year’s virtual ERI conference the DoD’s Ellen Lord, Under Secretary of Defense for Acquisition and Sustainment, told the summit that “US dependency on offshore suppliers is a growing national security risk.” Thus, after decades of moving chip fabrication, packaging, and testing capabilities offshore, the DoD is now proposing a step-by-step process for reconstituting the microelectronics supply chain in the USA.
They believe the development of a domestic microelectronics capability should be focused by segment including memory, logic, complex analog devices, and advanced packaging and assembly. Lord pointed out that while DoD does not drive the electronics market (constituting < 1 % of demand), they can drive significant R&D and have the power “…to convene thought leaders and establish public/private partnerships to provide a framework for growth.”
The major continuing concern for offshore supply is the ability of foreign entities to embed backdoors and/or malicious codes into our chips, which could result in loss of control of our critical DoD systems. While we can identify the technical path to ensure that all components are clean regardless of their manufacturing location, we need to have a domestic source for legacy and state-of-the-art products.
Lord points out that while we develop an enormous amount of microelectronic intellectual property here in the US, approximately 74% of that goes offshore to be produced and about 95% of it is packaged and tested offshore. The Govt, she contends, needs to partner with industry to begin “reshoring” some of that capability.
(IFTLE predicts this will become the next industry buzzword: “reshoring” )
One of the lessons of the current pandemic is that there are critical industries where we need to have an onshore presence like pharma and microelectronics. The consensus is that we can use the same defense production act that helped us just build ventilators to reshore our electronics industry.
It is believed that this becomes a reasonable economic concept because what may not have been sustainable 15-20 years ago, because of labor costs, is now much more sustainable due to the automation that has been applied to microelectronic manufacturing processes.
The DARPA Mission
Also at the 2020 ERI conference, Mark Rosker and Dev Palmer, Director and Deputy Director of DARPA’s MTO office respectfully, discussed DARPA’s mission (Figure 1), activities, and how they relate to the ERI.
Those of us with interest in advanced packaging are typically dealing with the Microsystems Technology Office (MTO). MTO’s mission is the development of high-performance, intelligent microsystems, and components.
A lot of you are at a loss as to who to talk to about DARPA proposals. Rosker presented this handy MTO program manager (PM) chart and offered some advice for those wishing to engage with the PMs in MTO (Figure 3).
“PMs are motivated by finding new and highly disruptive problems that can be the basis of new programs…Generally you should discuss ideas with them that might lead to new programs not present your approaches for ongoing programs.”
Palmer noted that “…the MTO mission of delivering microsystem capabilities to the DoD cannot be done if you cannot build the microsystems…. ERI helps MTO build closer relationships to the commercial semiconductor industry, which gives us better access to the semiconductor market.”
During the panel session, PMs were asked how small companies, startups, or university researchers can get access to the latest nodes when the inherent costs built into that now being over $0.5B at the 5nm node. The response was that the potential may have to be demonstrated on older nodes to get the costs down and then on the strength of the demonstrated capability, and market size arguments, try to find funding for the more costly node development. The PMs also offered that the use of standard chiplets, like those being developed in the DARPA CHIPS program, is focused on just this issue, i.e only using the latest expensive node where it is critical and reuse of other standard functions to lower the overall costs.
When asked about heterogeneous integration applications, Tom Kazior, recently PM from Raytheon, indicated that he was looking for “…the integration of three to five functions with the high density of silicon CMOS to get the best of both worlds” (Figure 3).