Commerce’s NSTC Strategy: An Unfinished Puzzle

What a new Commerce paper says — and doesn’t say — about “Centers of Excellence,” full-stack research, NSTC’s operating nonprofit, membership tiers, Manufacturing USA, and partner-nation consortiums

~Cover photo designed on Midjourney~

This article can also be found on the ChinaTalk newsletter.

When managers tell their team to design something from nothing, they must be prepared for a long, creative process — and inevitably they will receive countless progress reports detailing how most of the “puzzle pieces” are still laying on top of each other in a messy heap.

Earlier this week, the Department of Commerce gave its managers — Congress, and the American people — a progress report on the National Semiconductor Technology Center (NSTC). Described as an outline of Commerce’s “vision and strategy” for the NSTC, the report describes a few areas of fitted puzzle pieces — and a heap of fundamental and still-unsolved questions.

Fitted Puzzle Pieces

The substantive decisions that Commerce made in this progress report show that officials have been listening intently to recommendations from the US’s largest chip manufacturers. Dreamers in academia and fabless firms also got a bone.

Quietly Defining “Centers of Excellence” — Without Saying the Phrase

In the over 250 responses to Commerce’s January 2022 Request For Information (RFI) on how to implement the CHIPS Act, industry and academia respondents engaged in an intense tug-of-war in defining “centers of excellence.”

As discussed previously on Chip Capitols, a program like the NSTC could be fully centralized, fully distributed, or adopt a hybrid model. Universities, federal labs, and the most R&D-intensive semiconductor companies (namely, fabless design firms) prefer a strong central hub to coordinate the NSTC’s research centers. In contrast, chip manufacturers prefer decentralized centers of excellence with both specific focus areas and economically self-sustainable business models.

Commerce officials decided in this recent paper to give key US chipmakers what they wanted:

• There will be an Advanced Packaging Pilot Facility. Last year, Intel proposed setting up an Advanced Packaging Lithography Center and assisting with an Advanced Packaging Manufacturing Center. This facility would at least partially fulfill that wish. Critically, Intel advocated for a level of separation between the NSTC and the National Advanced Packaging Manufacturing Program (NAPMP) to ensure that spending on chip R&D would not detract attention from advanced packaging R&D. Commerce heeded Intel’s concern, saying that it “plans to establish NAPMP as a separate program within NIST.”

• There will also be a Heterogeneous Integration Technical Center. Micron last year urged the NSTC to create a Memory Center of Excellence that could enable next-generation 3D memory structure and heterogeneous integration. This facility would provide for much of Micron’s envisioned agenda; it would also conduct research on chiplets, which many stakeholders wanted.

Grand Challenges Should Cross Segments

Fundamental changes to semiconductor technology and supply chains require early orchestration at the research and development stage. Such plans involve coordinating the work of the NSTC’s various centers across industry segments (analog, memory, logic, etc.) in pursuit of tangible end-uses.

This Commerce report gave advocates for this full-stack research approach reason for hope. Officials said the NSTC’s grand challenges would be akin to past government projects like the DARPA Subterranean Challenge. Today’s grand challenges pursue tangible goals, such as lowering costs for semiconductor manufacturing and improving the environmental sustainability of chipmaking — and they stand in contrast to the siloed, incremental goals of particular chip industry segments.

Measurement on the Back-burner

The CHIPS Act charged the NIST with establishing a Metrology Research Program to research and coordinate work in measurement science, materials characterization, and the development of metrology tools for chip manufacturing.

After stressing the importance of these fields, the Commerce report subtly suggests that the NSTC’s metrology research will not receive independent funding. It says that the CHIPS Metrology Research Program will leverage “current programs and capabilities,” as well as build partnerships to promote the commercialization of metrology technology. The CHIPS R&D Office will also ensure that the NSTC and NAPMP sufficiently address “key metrology problems” — but this stance is a far cry from metrology having its own standalone research stream.

A Nonprofit Will Be Born

The US’s public-private partnerships (often called Federally Funded Research and Development Centers) have historically been run by nonprofits, so most observers believed that the NSTC would be as well — the key question was whether Commerce would select an existing nonprofit to run the NSTC.

Officials put any speculation to rest: the Commerce Department “anticipates the creation of a new, purpose-built, independent, nonprofit entity with the requisite neutrality, expertise, leadership, and capacity to serve as the operator of the NSTC.”

The Unfitted Puzzle-Piece Heap

Within the mountain of puzzle pieces yet to be fitted together, there is a group of pieces which must be sorted soon — and one over which officials may procrastinate.

Tiers of Membership

Congress was clear, RFI respondents were clear, and this recent report is clear: the NSTC will have a diverse membership.

As discussed previously on Chip Capitols, the NSTC will almost certainly have different tiers of membership to accommodate participants with less IP or cash to offer. Established semiconductor companies want lower-paying participants to receive fewer benefits from the consortium — a policy which would affect how broadly the NSTC could share its IP, facilities, and datasets among members.

The Commerce paper acknowledges this need, stressing that the NSTC’s “Design Enablement Gateway” will need flexible IP models. Different membership levels must receive different levels of access to the NSTC’s design licenses and data.

The key point, however, is that Commerce has identified only which pieces it needs to solve the “membership tier” puzzle — it has not begun fitting those pieces together.

Manufacturing USA — Why Are We Doing This Again?

I mean no disrespect to the authors of the CHIPS Act or their staff… Sometimes, however, lawmakers recall previously successful policy models while writing bills, then stick those ideas into a new bill without anyone asking them to do so.

And I think that’s what happened with the CHIPS Manufacturing USA institutes.

Manufacturing USA is a network of sixteen public-private partnerships in which industry, academia, and government cooperate to solve advanced manufacturing challenges. The CHIPS Act authorized up to three new Manufacturing USA institutes focused on advanced semiconductor manufacturing techniques and equipment.

Commerce’s recent report makes it clear that officials don’t yet know what to do with this line item: it simply said that its October 2022 RFI asked about “the design of, and requirements for, potential Manufacturing USA institutes.” Industry and academia replied with thoughtful insights to that RFI — but I have yet to find evidence that any significant players in the semiconductor ecosystem asked for Congress these institutes in the first place.

Partners and Allies, Theoretically, Are a Good Thing

There is little to say on this front: officials have not decided what role, if any, the NSTC’s foreign counterparts — like Beglium’s Imec or Taiwan’s ITRI — should play in its mission. The report says that the “NSTC will work to establish positive and synergistic relationships with partner nations and allies that share common values and objectives” — but these puzzle pieces definitely have yet to be sorted. Perhaps this could become a project for the US-EU Trade and Technology Council (TTC)?

Finishing the Puzzle

This vision and strategy report is not a final blueprint for the NSTC. It is a progress report — and should be judged as such.

Commerce Department officials have an enormous task ahead: to build an institution the likes of which the US has not built in decades. Critically, these officials will not be the ones to finish assembling the puzzle. Once Commerce finishes building the NSTC’s institutional skeleton, it will fall to its operating nonprofit and board of trustees to fit some of the smaller pieces together.

And how many puzzle pieces will Commerce leave unsolved for the board of trustees? That may be one of the most important questions for officials to answer.