Preparing US Workers for Advanced Manufacturing

As described in “Could an ARPA Help Resurrect US Manufacturing” (Issues, Winter 2025), William B. Bonvillian’s ARPA-M can help resurrect US manufacturing only if the nation’s workforce is prepared to take advantage of APRA-M’s innovations. Technologies such as artificial intelligence, big data, additive manufacturing, and flexible robotics will transform factories from labor intensive, low-skilled enterprises into automation intensive, high-skilled enterprises. Factory workers of the future will need to possess a skillset that falls in between a technician and an engineer, which Bonvillian classifies as a technologist. America must transform its vocational education system to train this new class of worker.

US vocational education needs to be dramatically reformed to lift it out of its current doldrums. The good news is that American students increasingly view vocational schools as an attractive alternative to four-year colleges. Policymakers must meet the moment and create high-quality vocational training with stakeholders in education, industry, and government. Germany’s vocational training system is critical to that nation’s manufacturing success, and is a system that US policymakers should work to re-create here. Fortunately, an example of this model is already working at a BMW plant in South Carolina. Students attend high school but are also paid to work in the BMW factory, allowing them to gain technical experience. Students are attracted to this program because of the pay and the path for future employment, BMW benefits from a pipeline of skilled workers to meet its labor needs, and the local community benefits from low youth unemployment. This type of vocational program should be replicated across America. This could help ameliorate the shortage of skilled workers at the semiconductor manufacturing company recently built in a Phoenix suburb by a Taiwan firm—an example of where a shortage of skilled workers impeded advanced manufacturing in America.

Community colleges, advanced manufacturing institutes, and manufacturing extension programs will also play a critical role in transforming vocational education into a lifelong process. Community colleges should be leveraged not only as an entry point for vocational training but as a place where workers can return throughout their careers to gain new skills. Wake Tech in North Carolina, which runs apprenticeship partnerships with Ford and other manufacturers, is an example of how a community college can serve as a link between employers and students. Advanced manufacturing institutes not only need to demonstrate the latest manufacturing technologies but also model to industry how investments in workforce education can provide lasting productivity improvements. Innovations from the institutes must reach both industry and the vocational education system, ensuring that US vocational training is teaching at the leading edge of innovation. Manufacturing extension programs could and should also play a critical role in demonstrating the importance of workforce education to small and medium-size firms.

ARPA-M’s innovations can be capitalized upon only if the American workforce is provided lifelong quality vocational training that is networked with the nation’s advanced manufacturing institutes and engaged with manufacturing extension programs.

Roy F. Richardson

Nicholas S. Vonortas

William B. Bonvillian provides a compelling assessment of some of the challenges facing the US manufacturing industry and the need for a mechanism to stimulate more manufacturing innovation. Bonvillian points to the Defense Advanced Research Project Agency (DARPA) as a potential model for a government agency that could fund high-payoff research for manufacturing and postulates the Department of Defense as a potential host organization for the envisioned ARPA-M.

Drawing on the characteristics of DARPA, Bonvillian describes three features for what an ARPA-M would look like: 1) be independent enough to do high-risk research, but linked to decisionmakers who can help transition technology to end users, 2) possess the resources and authorities to transition the technology into implementation, and 3) maintain a risk-tolerant culture that is required for an organization focused on breakthrough technologies.

The success of an ARPA-M within DOD will depend on its ability to deliver innovative manufacturing processes that are adopted to support defense needs, particularly by improving the performance or reducing the cost of defense systems, as well as spinning off commercial successes. To that end, I believe there are several additional technology transition considerations that must be addressed.

First, there must be an identified defense benefit for the adoption of the new manufacturing process. Successful transitions will more likely occur when filling an existing and identified gap in manufacturing capability. Second, active and early collaboration with potential transition partners must be a feature of every funded project. A complicating factor here is that, unlike with DARPA, the relevant end users for defense manufacturing technologies are typically private-sector companies along with their defense acquisition program partners in the military services. One way to address this challenge would be to partner ARPA-M with the Manufacturing USA network as well as other relevant consortia of private-sector end users. Lastly, ARPA-M projects should have clearly defined technical goals to outline the tracking of desired capabilities, technical milestones, and required funding. A popular critique of DARPA is that it does not consistently define and track its transition outcomes, thereby complicating its ability to effectively improve transition processes. An ARPA-M must have clear and consistent metrics and data collection processes for transitioning new technologies and tracking their use for defense and other sectors.

Ultimately, the success of ARPA-M will be determined by the success of its program managers. These individuals should be given resources, such as formal training, on how to facilitate technology transition of their projects through engagement with the defense acquisition community and its supporting industrial base. Training should include discussions of past examples of successful transitions, an understanding of flexible procurement, and budgeting authorities that can streamline processes as well as external factors—including global economic conditions, congressional action, and public sentiment—that can play a role in project success.

The United States’ national security strength is often credited to innovation and technological achievement. None of the technological advances that our national security depends on today would be possible without the contributions from a robust defense manufacturing base. We must redouble our efforts to inject a spirit of innovation into this critical part of our national security enterprise.

Arun Seraphin

Today’s competition for economic and geopolitical leadership is driven by the strength of a nation’s innovation system. This system consists of an interlinked network of firms, universities, banks, government departments, regulatory agencies, venture capital funds, technical standards institutions, schools, and other providers of education and training, to name but a few.

Manufacturing lies at the core of this innovation system. However, the United States has for decades offshored manufacturing and disinvested in the workforce and infrastructure needed to support a vibrant manufacturing sector, and this erosion represents a significant vulnerability in the nation’s strategic competition.

What can help resurrect US manufacturing? William B. Bonvillian argues that standard policy approaches focused on tax, trade, and currency valuation are not adequate. What is needed are institutions that connect manufacturing firms to existing networks in research and development, financing, capabilities in workforce training, and government procurement.

Establishing and maintaining productive connections across the innovation system requires a suite of bridging institutions that address the cooperation challenges unique to each of these relationships. For example, manufacturers that compete in a given marketplace also need to cooperate with each other and with regional vocational colleges and state and local governments to train the skilled technical workforce.

For many policymakers, ARPA stands for the generic partnership instrument that addresses these bridging needs. Based on the success of the Department of Defense’s DARPA, the collection of variants already includes ARPA-Health, ARPA-Energy, and APRA-Infrastructure. DARPA itself is successful because it is tailored to connect key actors across the innovation system to solve specific technical challenges. It does so within a specialized organization that is autonomous, small in size, and features limited tenure of its program managers.

An ARPA-Manufacturing that can connect manufacturers to agency missions and that commands the resources to transition new technologies into implementation can be a valuable tool. But it is not a panacea. ARPA-M has to be tailored to the specific collective action challenges faced across the interface between manufacturing and the other gears in the innovation system. As Bonvillian points out, “An ARPA for manufacturing will be successful only if it is one component of a larger ecosystem that addresses research, policy, finance, industry collaboration, and worker education.”

Sujai Shivakumar

William Bonvillain makes many excellent points and articulates the slow-but-steady erosion of US manufacturing capability. But his observations and recommendations, especially regarding the offshoring of manufacturing, don’t hold up. I say this as an engineer who spent his career at the intersection of science and economics, including decades in the commercial microelectronics industry and a term as program manager at the Defense Advanced Research Projects Agency (DARPA), the original ARPA within the Department of Defense.

Famously, for example, semiconductor manufacturing companies have eschewed growing US manufacturing in favor of international expansion. But is it, as Bonvillian implies, lack of ideas or expertise that could be remedied by ARPA-style innovation in manufacturing technology? Perhaps not, because many of the companies expanding overseas have significant US manufacturing already. They chose to move manufacturing elsewhere, initially exporting the manufacturing expertise with US ex-pats who run the facility. This implies there is something inherently unattractive about manufacturing expansion in the United States.

Any manufacturing scale-up must clearly deliver economic benefit to the company. This benefit must be weighed against the risks. Not all risks can be mitigated by the company. Depreciation rule changes have significantly lowered the return on investment of microelectronics manufacturing. Taxation encourages leaving offshore income offshore as well, where it must be spent offshore, a vicious cycle. In this regard, the United States is clearly in competition with other geographies, those that are more generous from an accounting and taxation perspective. It would behoove architects of resurgent US manufacturing to consider complementary financial rules along with such ideas as artificial intelligence and robotic assembly. Otherwise, sensible CEOs will expend their time, capital, energy, and reputation on less risky endeavors.

So can a government ARPA program work? If all it does is deliver some technological solutions, the answer is no. Further, government is historically a poor partner: it is frivolous, changing with the political winds. Even DARPA has a poor track record of delivering success at the commercial level. Of course, its projects are connected with military applications, so commercial success is not essential. But failure is more deeply rooted than that. My observation inside DARPA was that some considered commercial success a negative result. It implied industry would have done the work anyway, so government should not have funded the program. Regardless of the reason, organizational resistance to commercial exploitation of government funding generates a strong headwind to any successful government involvement in manufacturing resurgence.

As another detriment, the government has spent decades squeezing US manufacturing with regulations—environmental, financial, social, and political—making it costlier and riskier. CEOs are not stupid: they see government siphoning off profits their foreign competitors don’t lose, then they see government offering high-burden, socially manipulative projects to get some of that money back for a few years. Will private industry participate in such a risky new government program? Perhaps. Will they commit? Not likely.

Yes, it hurts to hear it, but the United States is not a stable place to invest in manufacturing. Fix that, and it is likely the other problems will evaporate. This is not an argument to ignore the issues Bonvillian addresses, but to understand the broader issues. Of course, it will be a long, slow process. It will take decades for the commercial sector to regain trust in the US environment. And perhaps the creation of a few trillionaires, those who accepted the enormous risks and proceeded anyway.

Steve Trimberger

William Bonvillian’s valuable analysis describes a fundamental challenge facing the American economy. He rightly emphasizes the importance of manufacturing for innovation, growth, and national security, and importantly, the well-paying jobs needed to maintain political stability. The article is also very timely given the massive rise in East Asian competition across multiple sectors. His informed references to the challenges facing the US defense industry, and the nation’s inability to scale up production to restore its defense industrial stockpile and meet the needs of the Ukrainian conflict, underscore the importance of the topic and his recommendations.

His policy prescription to establish an ARPA for manufacturing (ARPA-M) is excellent. It is a proven model both within the Department of Defense, with its Defense Advanced Research Project Agency (DARPA), and within the Department of Energy, with its Advanced Research Projects Agency-Energy (ARPA-E), an institutional concept advanced in no small part through Bonvillian’s own efforts. His recommendations for where to locate an ARPA-M are also right on the mark. The National Institute of Standards and Technology clearly has a manufacturing vocation, but it also is a deeply risk-adverse institution, lacks scale in resources, and faces significant procurement needs. While locating an ARPA-M within the Defense Department as a standalone entity might make sense given that its focus will be substantially different from DARPA, it would be safer as a major program enveloped in DARPA where it would be less likely to fall victim to errant congressional budget cutting.

Bonvillian’s suggestion to provide tax incentives to attract longer-term venture funding is also important. Creatively applied, such incentives could address a major gap in US funding for manufacturing—as would complementary programs to provide federal matching funds for investors in early-stage manufacturing companies, a space that venture funds often neglect given the longer lead-times and higher upfront costs associated with manufacturing. For industry’s importance to our defense capabilities, we need to begin to think more creatively to include loans, loan guarantees, cost-plus contracts, and capital infusions. We also need to encourage more rapid acquisition through much greater use of Other Transaction Agreements (OTAs) to bring new high-tech products, such as advanced drones, to the warfighter faster and cheaper.

We need to up our game in terms of resources and drop old bromides about “picking winners and losers” if we are to even begin to level the playing field with China. This holds for the author’s observation about the effectiveness but lack of scale of the Manufacturing USA’s network. The member institutes are promising but their budgets are too small and the funding horizons too short to fulfill their mission. Their resources pale in comparison to the German Fraunhofer Institutes on which they were modeled with their long-term funding and assured capital budgets. Fundamentally, if we are to address the reality of China’s efforts in this space we need to think outside the box. If the United States is to compete, we need to get real both in terms of focus, procedures, and resources. Bonvillian should be heard.

Charles Wessner