Arsenal of Democratic Cognition

In 1940, as Nazi Germany and Imperial Japan marched across Europe and Asia, the United States faced a crisis of production. The challenge was not simply to build higher-quality tanks or planes, but to construct an entire industrial ecosystem capable of meeting wartime demands and overwhelming adversaries who possessed first-mover advantages and immediate battlefield superiority. [1] To accomplish this task, President Franklin Roosevelt did not nationalize the factories of Detroit, nor did he order the Army to seize General Motors. Instead, his administration forged a ruthlessly efficient and uniquely American weapon: a public-private partnership that leveraged the might of free enterprise and government into the famous “Arsenal of Democracy.” 

The United States now confronts an analogous issue of mobilization in the race towards Artificial General Intelligence (AGI). China, the United States’ primary strategic adversary, has long recognized the strategic implications of AGI and, since 2017, has used state policy to pursue research and development through coordinated institutional frameworks—from the Beijing Institute for General Artificial Intelligence to municipal five-year plans. [2]  Accordingly, the American advantage is eroding at an especially alarming pace. Chinese frontier models have closed the benchmark performance gap with their American counterparts from years to mere months. [3] Meanwhile, Chinese open-weight models command a significant growing share of global AI usage, and in 2025 Beijing stood up a trillion-RMB ($138 billion) National Innovation Investment Guidance Fund—a twenty-year commitment targeting frontier AI domains. [4][5] The United States retains its lead in raw compute capacity and private investment, but neither advantage is self-sustaining. For the United States, therefore, AGI development is a matter of supreme national interest.  

The launch of the Genesis Mission in November 2025 marked a pivotal recognition by the Trump administration that the race for AGI is the geopolitical center of gravity for the twenty-first century. The Genesis Mission is a national initiative led by the Department of Energy that aims to leverage the nation’s seventeen National Laboratories, federal supercomputing resources, and vast scientific datasets to build an integrated AI platform for accelerating scientific discovery. Tellingly, the executive order authorizing the mission frames it as “comparable in urgency and ambition to the Manhattan Project”. [6] Yet the Genesis Mission, by design, is a government-lab scientific platform. It does not provide a mechanism for integrating private-sector scaling, commercial innovation, or the massive industrial coordination required across firms. Thus, to secure American geopolitical and technological primacy, the United States must elevate this initiative into a national project—a twenty-first-century public-private partnership for AGI development, harnessing the scaling power of the federal government and the unmatched innovation of the private sector. [7]  

What AGI is—and Why The Distinction Matters

AGI remains conceptually contested, but can be broadly defined as an AI system that demonstrates human-level or superior performance across a wide range of cognitive tasks without relying on domain-specific training. [8] Rather than excelling at a single function—such as image recognition or translation—AGI would display an ability to generalize and possess a degree of common-sense reasoning across diverse domains. [9] This definition is particularly important because it distinguishes AGI from current advanced, yet narrow AI systems. Large language models like GPT-5 or Claude excel at generating human-like text and assisting with complex reasoning; however, they are still fundamentally specialized tools optimized for particular tasks via large-scale pattern recognition. AGI systems, by contrast, are capable of autonomous goal-directed behavior, cross-domain problem-solving, and possibly iterative self-improvement. [10]  

If such a capability is realized, its strategic significance is clear: these systems could conduct parallelizable research, accelerate scientific and engineering discovery, optimize industrial processes, and support complex military and intelligence workflows at speeds and scales that far outpace human teams. That is why the central policy question is not whether AGI will matter—frankly, it already does—but whether the United States can build and secure the enabling ecosystem fast enough before adversaries do. 

Why Neither Markets Nor The State Can Do It Alone

The enormous task of researching and developing AGI comes with immediate physical constraints. In this regard, American primacy relies no longer on steel or rubber, but on compute and energy. Modern AI data center clusters now face interconnection wait times extending up to seven years in key jurisdictions, with power requirements reaching gigawatt-scale per campus. [11] AI workloads demand 70 to 80 kilowatts per rack compared to 8 to 15 kilowatts for traditional data centers, with modern GPU clusters consuming 700 to 1,200 watts per processor versus 150 to 200 watts for conventional server CPUs. [12] Goldman Sachs Research projects data center power demand will grow 165 percent by 2030, driven predominantly by AI workloads. [13] The required capital alone is staggering. OpenAI projects spending $115 billion between 2024-2029. At the global level, data center infrastructure for AI alone demands $5.2 trillion by 2030, roughly the GDP of Germany. [14][15] Meanwhile, energy infrastructure requires an additional $1.4 trillion over five years, all while talent shortages affect 75 percent of AI companies. [16][17] 

AGI also requires convergence across multiple technical frontiers simultaneously. No single firm has the incentive or capacity to solve these coordination problems cooperatively. The free market, while an engine of innovation, lacks the structural capacity to absorb the volatile variables involved: regulatory and market uncertainty, grid upgrades, transmission corridors, nuclear licensing pathways, and accelerated interconnection queues require cross-jurisdictional authority, long time horizons, and credible commitments that exceed what private firms—even the most capable—can reliably sustain. [18] The strategic externalities—deterring adversaries, preserving technological leadership, and preventing systemic vulnerability—cannot be solved in a single firm’s ROI model. Yet the path to victory lies not in mimicking our adversaries’ state-directed authoritarianism. China’s civil-military fusion model coerces private compliance under party direction. [19] The American answer must be different, rooted in democratic tradition and tested historical precedent.  

The American Tradition Of Directed Public-Private Partnership

Every major public-private partnership in American history has been forged in the crucible of a geopolitical crisis, specifically at moments when a near-peer or peer state exposed a critical vulnerability in the nation’s technological or industrial base. In 1940, it was the Axis powers’ battlefield superiority and the realization that the United States lacked industrial capacity to wage a two-front war in the Pacific and Europe. In 1957, it was the Soviet launch of Sputnik, a technological surprise that shattered American confidence in its own scientific superiority and exposed the absence of an institutional mechanism for sustaining technological advantage against a peer adversary. [20] And in the mid-1980s, it was Japan’s capture of dominant market share in semiconductor manufacturing that triggered alarm across the national security establishment—a threat that the 1987 Defense Science Board report made explicit, warning that American dependence on foreign-produced microelectronics jeopardized the precision-strike capabilities underpinning U.S. military doctrine in a hypothetical conflict against the Soviet Union. [21] In each case, the catalyst saw the United States confronting a strategic competitor that had seized the initiative in a domain critical to the balance of power, and the existing institutional arrangements proved inadequate to the scale of the response required.  

What followed in each instance was a deliberate construction of a partnership calibrated to the nature of the strategic threat. President Roosevelt’s War Production Board (WPB) placed industry executives in leadership positions—Sears’ Donald Nelson as chairman, General Motors’ William Knudsen as production coordinator—and relied on a contract-based model in which the government financed facilities through government-owned, contractor-operated arrangements and guaranteed procurement, while private firms retained operational control. [22]  

The Defense Advanced Research Projects Agency (DARPA), created in 1958 in the wake of Sputnik, refined the partnership model into a durable institutional form, maintaining no laboratories of its own and operating with a deliberately skeletal bureaucracy; its program managers—empowered to define research objectives and select performers through competitive contracts with private firms and universities—serve fixed terms of three to five years, preventing bureaucratic entrenchment and ensuring that the agency’s portfolio evolves with the threat environment rather than calcifying around institutional interests. [23][24] Semiconductor Manufacturing Technology (SEMATECH) adopted a consortium structure: fourteen U.S. semiconductor firms contributed half the annual $200 million budget, the Department of Defense matched the other half through DARPA, and member companies—not government officials—set the research agenda. [25] The consortium was explicitly prohibited from selling semiconductor products, preserving competitive dynamics among its own members. [26] In every case, government and industry remained distinct spheres with varied interests, objectives, and constraints—but the partnership between them produced output that neither could have generated alone.  

The results speak to the model’s efficacy as an instrument of geopolitical competition. The WPB directed the production of approximately $185 billion in armaments and supplies, converting roughly 90 percent of U.S. manufacturing to war production by 1944 and enabling aircraft output to surge over 190,000 in less than four years—industrial outputs that proved decisive in overwhelming the Axis powers. [27][28] DARPA’s portfolio has produced transformative technologies—the Internet, GPS, stealth aircraft, precision-guided munitions, voice recognition, and early mRNA vaccine research—that have reshaped the military balance while generating trillions of dollars in civilian economic value [29]. Today, approximately 70 percent of DARPA’s programs incorporate some form of artificial intelligence, machine learning, or autonomy [30]. SEMATECH’s impact was more targeted but no less significant: by the early 1990s, the U.S. semiconductor industry had reclaimed global leadership—a turnaround attributed directly to SEMATECH, which closed the quality gap with Japanese producers, drove up manufacturing yields, compressed product cycles, and rebuilt a quality domestic equipment base [31]. The government contributed approximately $870 million over the partnership’s life while member companies contributed $863 million, realizing more than $34.7 billion in tax revenue [32]. 

Equally important is what happened after the imperative had shifted where these institutional models existed, not as permanent bureaucracies but as tactical tools. The WPB dissolved in November 1945, and its remaining functions transferred to civilian administration [33]. On the other hand, DARPA endured because the threat environment endured: the agency’s mission of preventing strategic technological surprise has remained relevant across the Cold War, the post-Cold War era, and the return of great-power competition. SEMATECH’s board voted in 1994 to decline further federal funding after fiscal year 1996—a voluntary sunset that demonstrated the model’s self-limiting character [34]. The common logic is self-evident—the government identifies a strategic gap and then absorbs the downside risk while clearing the coordination barriers that private firms cannot overcome alone. The private sector retains ownership and operational autonomy. And when the partnership has served its strategic purpose, it dissolves or evolves—rather than metastasizing into a permanent bureaucratic fixture. This kind of industrial policy is a proven instrument designed and deployed at moments of maximum consequence.  

Answering The Critics

A serious proposal of this proportion must confront, rather than dismiss, its strongest objections. Three distinct critiques merit thorough engagement. Libertarian objections posit that government intervention distorts markets by picking winners and losers. Moreover, policymakers lack the knowledge necessary to allocate capital efficiently identified as the knowledge problem [35]. This critique carries genuine force in the general case. However, it conflates two distinct government functions: directing innovation and enabling the conditions for innovation. The partnership model advanced here does not ask the government to choose which firm will build AGI or which product will dominate the market; rather, it asks the government to do what only the government can do: clear regulatory bottlenecks in energy permitting and grid interconnection, underwrite the infrastructure risk that private capital markets cannot price over sufficient time horizons, and guarantee demand through procurement commitments. These are simply preconditions for competition. In particular, DARPA’s seven decades of sustained output demonstrate that the government can fund high-risk research without directing commercial outcomes—its program managers define problems while the competitive contract structure ensures that market actors, not bureaucrats, determine which approaches succeed [36].  

The second critique comes from the conservative institutionalist camp, which accepts the premises of strategic competition but remains skeptical that the federal government can execute industrial policy without further introducing inefficiency and waste. The skepticism is certainly well-founded—the implementation of the CHIPS and Science Act has already demonstrated the risks of regulatory accretion, with progressive policy add-ons and existing permitting requirements slowing disbursement and inflating costs [37]. The answer, however, is to design it correctly rather than abandon it altogether. The American historical record provides a template. SEMATECH succeeded in part because it was industry-led and provided fixed terms which ensured that the agency’s portfolio remained responsive to evolving technological realities rather than institutional inertia. The WPB dissolved entirely once the war ended. Effective partnership design requires structural disciplines: competitive procurement, milestone-based funding, independent auditing, and built-in sunset provisions that force periodic reassessment of whether the rationale of the institution still holds. 

The third and most serious critique draws on public choice theory: any large-scale government program will attract rent-seekers who capture the benefits for themselves at the expense of the public interest. Many have documented this pattern extensively, noting that industrial policy programs tend to benefit the few at the expense of the many and that lobbying dynamics predictably distort allocation decisions [38]. This is the objection that demands careful elaboration, because the risk is real and historically recurrent. The mitigation lies in institutional design that makes capture innately difficult. 

SEMATECH’s charter required independent commercial audits of all government-funded expenditures, and the Secretary of Defense retained oversight authority throughout the partnership’s life [39]. DARPA’s flat organizational structure, rapid program cycling, and reliance on empowered individual program managers—rather than standing committees vulnerable to industrial lobbying—limit the stable relationships through which capture typically operates [40]. Competitive procurement, in which multiple private firms compete for contracts on the basis of technical merit rather than political connections, further limits rent-seeking. Of course, no institutional design eliminates these risks entirely. Yet, the question is whether the risks of partnerships are greater than the risks of inaction in the face of a strategic competitor that shares no such inhibitions.  

A National Agi Consortium

The Genesis Mission supplies a scientific platform but does not supply the partnership required to lay the foundations for a project of such enormous proportions. Its instrument is the national laboratory network, with its purpose to accelerate research. The functions that lie beyond its mandate—related to infrastructure and financing—belong to a different institutional tradition. The WPB, DARPA, and SEMATECH addressed precisely this set of problems in their eras; subsequently, the establishment of a National AGI Consortium—a federally chartered, industry-led body modeled on the institutional architecture of the War Production Board, structural disciplines of SEMATECH and the operational philosophy of DARPA—would close the gap between scientific capacity and material deployment. A scientific platform without an industrial ecosystem yields breakthroughs that never reach scale; an industrial ecosystem without a scientific platform yields mobilization without substance. Each requires the other. Because the Genesis Mission already exists, the remaining bulk of the work focuses itself on the architecture—the assembly of partnership institutions around capabilities Congress has already authorized. 

The two institutions are complementary by design. The Genesis Mission provides the technological core while the Consortium provides the necessary conditions under which a technological core becomes material national capability. The Consortium’s mandate should be threefold: first, coordinating the infrastructure preconditions for AGI development, including energy permitting, grid interconnection, and shared compute access through the Department of Energy national laboratory network established by the Genesis Mission; second, funding pre-competitive research in areas where no single firm has sufficient incentive to invest alone, such as evaluation standards, interpretability, and alignment benchmarks; and third, maintaining competitive procurement structures that prevent any single firm from monopolizing federally supported resources. Governance should rest with a joint board holding a private-sector majority, with member firms contributing matched funding alongside federal investment disbursed on a milestone basis. Critically, the consortium must not pick commercial winners. Like SEMATECH, it should be prohibited from developing proprietary AGI products with its primary function to raise the floor of the enabling ecosystem. Sunset provisions are non-negotiable: a five-year congressional reauthorization cycle would force periodic reassessments, preventing institutional calcification. Independent auditing and Department of Defense oversight, as in SEMATECH’s charter, would provide accountability without micromanagement. Of course, this is merely a rudimentary framework that must logically be fleshed out. Yet the blueprint lies in historical precedent to which policymakers and private industry alike must harken back to.  

However, a national project of this scale cannot survive on technical merit alone; it endures only when it sits atop a durable political coalition. In the late nineteenth century, protective tariffs and continental rail grants survived populist agitation for two generations because their authors understood this and acted on it. They cultivated the congressional constituencies whose interests their policies served, binding industrial development to the political base that would defend it through successive electoral cycles [41]. The Consortium would function analogously. Its three operational mandates each map onto an identifiable constituency: infrastructure coordination engages the utility states and grid-belt delegations whose districts will host the gigawatt facilities and transmission corridors that frontier compute demands; procurement engages the defense-industrial geography that already organizes congressional behavior around the Pentagon’s long-cycle programs; pre-competitive research engages the network of research universities and national laboratories distributed across most states. Hypothetically, the Consortium’s political durability would derive from this distribution. The five-year reauthorization cycle is therefore a feature, the mechanism by which the coalition is periodically reaffirmed and the project advanced. 

Conclusion

The race for AGI has become a concrete strategic challenge with a defined adversary, measurable stakes, and a rapidly closing window of advantage. The United States possesses a rich legacy of historical models and the legislative foundation to act. The Genesis Mission has already been authorized. This is no longer a debate about the merits of industrial policy in the abstract. There should be no doubt that China intends to displace the United States as the undisputed technological leader. What remains is the will to do so: the decision to build, around an existing scientific platform, the industrial ecosystem that transforms capability into national power. 

The American tradition does not guarantee success but supplies a method, primarily through the partnership model that has defined the country’s most consequential moments of strategic mobilization. From the WPB to SEMATECH to DARPA, the United States has repeatedly demonstrated that the federal government and the private sector can achieve together what neither can accomplish alone—provided the partnership is designed with structural disciplines that preserve competition, limit capture, and most importantly, enforce accountability. 

What is at stake warrants clarity. In 1940, the Arsenal of Democracy was a mobilization of industrial capacity—consequential, historic, and ultimately reversible. The crisis today is different in form but identical in structure: a strategic adversary with first-mover ambitions, a technological race with civilizational implications, and a challenge that exceeds the capacity of any single actor. The nation that consolidates this infrastructure first sets the terms of the century that follows. The United States has built decisive advantages before under conditions of maximum pressure and minimum time. It must do so again, forging an Arsenal of Democratic Cognition—a partnership worthy of the tradition that built it, and equal to the challenge that demands it. H

Benson Pham ’26 served as Chapter Founder and President of the AHS chapter at San Diego State University, where he majored in Political Science and Government.

---

NOTES:

[1] Alan L. Gropman, “Mobilizing U.S. Industry in World War II: Myth and Reality,” McNair Paper 50. Washington, DC: Institute for National Strategic Studies, National Defense University, August 1996. https://apps.dtic.mil/sti/tr/pdf/ADA316780.pdf. 

[2] Matthew Johnson, “AGI Has Quietly Become Central to Beijing’s AI Strategy,” China Brief, October 1, 2025, https://jamestown.org/agi-has-quietly-become-central-to-beijings-ai-strategy/.

[3] “China’s AI in 2025: Progress, Players, and Parity,” Venturous Group, October 27, 2025, https://www.venturousgroup.com/resources/chinas-ai-in-2025-progress-players-and-parity/.

[4] “Two Loops: How China’s Open AI Strategy Reinforces Its Industrial Dominance,” U.S.-China Economic and Security Review Commission, March 22, 2026, 

https://www.uscc.gov/sites/default/files/2026-03/Two_Loops–How_Chinas_Open_AI_Strategy_Reinforc es_Its_Industrial_Dominance.pdf. 

[5] “Measuring the US-China AI Gap,” Recorded Future, May 7, 2025, https://www.recordedfuture.com/research/measuring-the-us-china-ai-gap.

[6] United States, Executive Order of the President, Executive Order 14363, “Launching the Genesis Mission,” November 24, 2025, https://www.whitehouse.gov/presidential-actions/2025/11/launching-the-genesis-mission/.

[7] U.S.-China Economic and Security Review Commission, 2024 Annual Report to Congress, 118th Cong., 2nd sess., November 2024, https://www.uscc.gov/annual-report/2024-annual-report-congress.

[8] Meredith Ringel Morris et al., “Levels of AGI: Operationalizing Progress on the Path to AGI,” Google DeepMind, November 4, 2023. 

[9] “What Is Artificial General Intelligence (AGI)?,” Google, https://cloud.google.com/discover/what-is-artificial-general-intelligence. 

[10] “What is Artificial General Intelligence (AGI)?,” IBM, https://www.ibm.com/think/topics/artificial-general-intelligence. 

[11] Martin Stansbury et al., “Can US Infrastructure Keep Up with the AI Economy?,” Deloitte Insights, June 24, 2025, https://www.deloitte.com/us/en/insights/industry/power-and-utilities/data-center-infrastructure-artificial-in telligence.html. 

[12] “Power Requirements for AI Data Centers,” Resilient Infrastructure Hanwha Data Centers, September 18, 2025, https://www.hanwhadatacenters.com/blog/power-requirements-for-ai-data-centers-resilient-infrastructure. 

[13] Carly Davenport et al., “AI Is Poised to Drive 160% Increase in Data Center Power Demand,” Goldman Sachs, May 14, 2024, https://www.goldmansachs.com/insights/articles/AI-poised-to-drive-160-increase-in-power-demand. 

[14] Jack Houghton, “The Real Cost of AGI—According to OpenAI,” In The Loop (podcast), Episode 30, Mindset AI, accessed January 7, 2026,  https://www.mindset.ai/blogs/in-the-loop-ep30-the-real-cost-of-agi. 

[15] Jesse Noffsinger, Mark Patel, and Pankaj Sachdeva, “The Cost of Compute: A $7 Trillion Race to Scale Data Centers,” McKinsey & Company, accessed January 7, 2026,  

https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-cost-of -compute-a-7-trillion-dollar-race-to-scale-data-centers. 

[16] “2025-2030: $1.4 Trillion in Energy Infrastructure Opportunities,” LandGate, accessed January 7, 2026,  https://www.landgate.com/news/2025-2030-1-4-trillion-in-energy-infrastructure-.opportunities. 2026, https://www.landgate.com/news/2025-2030-1-4-trillion-in-energy-infrastructure-opportunities. 

[17] “How Policies Can Help Solve Talent Shortages in Critical Sectors,” Axios, accessed January 7, 2026, https://www.axios.com/sponsored/how-policies-can-help-solve-talent-shortages-in-critical-sectors. 

[18] Houghton, “Real Cost of AGI.”, Mindset AI, September 9, 2025. https://www.mindset.ai/blogs/in-the-loop-ep30-the-real-cost-of-agi.

[19] U.S. Department of State, “The Chinese Communist Party’s Military-Civil Fusion Policy,” January 13, 2020, https://2017-2021.state.gov/military-civil-fusion. 

[20] “About DARPA,” Defense Advanced Research Projects Agency, https://www.darpa.mil/about.

[21] “SEMATECH Revisited: Assessing Consortium Impacts on Semiconductor Industry R&D,” in National Research Council, Securing the Future: Regional and National Programs to Support the Semiconductor Industry (Washington, DC: The National Academies Press, 2003). 

[22] “What Does World War II Teach Us About Industrial Policy Today?” Factory Settings, March 2026, https://www.factorysettings.org/p/what-does-world-war-ii-teach-us-about. 

[23] “Defense Advanced Research Projects Agency (DARPA),” Encyclopedia Britannica, February 3, 2026, https://www.britannica.com/topic/Defense-Advanced-Research-Projects-Agency.

[24] William B. Bonvillian and Richard Van Atta, “The DARPA Model for Transformative Technologies,” in William B. Bonvillian et al., eds., The DARPA Model for Transformative Technologies (Cambridge: Open Book Publishers, 2019).

[25] “U.S. Semiconductor Manufacturing: Industry Trends, Global Competition, Federal Policy,” Congressional Research Service, R44544, https://www.congress.gov/crs-product/R44544. 

[26] Douglas A. Irwin and Peter J. Klenow, “Sematech: Purpose and Performance,” Proceedings of the National Academy of Sciences 93 (July 1996): 12739–12742, https://pmc.ncbi.nlm.nih.gov/articles/PMC34130/.

[27] Aaron Purcell, “War Production Board,” EBSCO Research Starters, 2021, https://www.ebsco.com/research-starters/military-history-and-science/war-production-board. 

[28] University of Warwick, “Annual Number of Combat Aircraft Produced by the Major Powers During the Second World War from 1939 to 1945,” Statista, January 1, 1998, https://www.statista.com/statistics/1336929/wwii-combat-aircraft-production-annual/. 

[29] “About DARPA,” Defense Advanced Research Projects Agency.

[30] David Vergun, “DARPA Aims to Develop AI, Autonomy Applications Warfighters Can Trust,” U.S. Department of War, March 27, 2024, 

https://www.war.gov/News/News-Stories/Article/Article/3722849/darpa-aims-to-develop-ai-autonomy-ap plications-warfighters-can-trust/. 

[31] Charles Wessner and Thomas Howell, cited in “Sematech: A Public-Private Partnership for Spurring Domestic Manufacturing,” Bipartisan Policy Center, February 2024, https://bipartisanpolicy.org/wp-content/uploads/2024/02/Sematech-A-public-private-partnership-for-spurr ing-domestic-manufacturing.pdf. 

[32] “SEMATECH 1987–1997: A Final Report to the Department of Defense,” February 21, 1997, https://www.esd.whs.mil/Portals/54/Documents/FOID/Reading%20Room/Science_and_Technology/10-F -0709_A_Final_Report_to_the_Department_of_Defense_February_21_1987.pdf. 

[33] “War Production Board,” EBSCO Research Starters. 

[34] “U.S. Semiconductor Manufacturing,” Congressional Research Service. 

[35] Friedrich A. Hayek, “The Use of Knowledge in Society,” American Economic Review 35 (September 1945): 519–530; cited in “Industrial Policy, Whether Republican or Democrat, Is Anti-Liberty,” Cato, January 4, 2025, https://www.cato.org/blog/industrial-policy-whether-republican-or-democrat-anti-liberty.  

[36] Bonvillian and Van Atta, “The DARPA Model for Transformative Technologies.”

[37] Scott Lincicome, “Scott Lincicome’s Memo to Industrial Policy Proponents: Political Reality Isn’t Optional,” The Dispatch, April 2023, cited in Cafe Hayek, 

https://cafehayek.com/2023/04/scott-lincicomes-memo-to-industrial-policy-proponents-political-reality-is nt-optional.html. 

[38] Scott Lincicome, cited in “Industrial Policy 101,” Medium, May 2022, https://rzadek.medium.com/industrial-policy-101-a1421030ef5c. 

[39] “SEMATECH 1987–1997: A Final Report to the Department of Defense.” 

[40] Bonvillian and Van Atta, “The DARPA Model for Transformative Technologies.” 

[41] Richard Franklin Bensel, The Political Economy of American Industrialization, 1877–1900 (Cambridge: Cambridge University Press, 2000), 457–509. 

Image: “Virginia Tech – data center” by Christopher Bowns, retrieved from https://commons.wikimedia.org/wiki/File:Virginia_Tech_-_data_center.jpg. This file is licensed under the Creative Commons Attribution-Share Alike 2.0 Generic license.