Rebuilding the American Offer: Manufacturing, Statecraft, and the Legislative Package Industrial Renewal Requires

America’s strategic position rests on two compounding advantages: what America can do, and what its allies and partners cannot afford to lose. The military guarantee, which includes open sea lanes, forward deployment tripwires, assurance of competence, and the nuclear umbrella, creates a security dependency.1 The economic value proposition, consisting of capital access and credit, investable markets, shared standards, and competitive products at fair prices, creates the commercial one. Together, they have constituted an offer no other nation could match.

Today, one leg is failing and the other is limping. Economic statecraft requires resilient supply chains and near competitive products. The United States possesses every ingredient needed to dominate economic statecraft except the one that makes the others deployable: a resilient, competitive, broad-based manufacturing sector. America rebuilding its manufacturing base is not about outcompeting China on price. China’s purposeful, subsidized overinvestment-with-low-expected-returns strategy will nearly always win in the world marketplace, where price is the main point of competition.

Many countries would prefer to trade, venture, and work with America, however. And many would likely be willing to pay a little more to do so.2 But America just fell off the bandwagon of leading-edge and non-leading-edge manufactured goods across its entire supply chains.

America must rebuild its manufacturing to execute that portion of economic statecraft where goods, trade, joint manufacturing, and IP are part of its overall country relationships. America must remove chokepoints and dependencies where it is at risk. And America needs manufacturing to take advantage and thrive in new technological frontiers, as AI delivers tremendous opportunity for embodied-AI products, medicines, materials, and applocations we don’t yet know.

Can America build, over time, a virtual trading bloc of aligned countries? Is that a good idea or even possible? That’s a longer-term question that’s literally not feasible today, given America’s current dependence on China.

This is the Bedrock trap: feeling wealthy because you are, while losing the ability to act because your strengths have no physical expression in the world.3 By every major measure, the American economy remains the envy of the G7, with leading growth, deep capital markets, the world’s reserve currency, the strongest alliance network in history, and unmatched (though challenged) innovation capacity. These are real assets. They are also, without broad-based competitive manufacturing, increasingly undeployable as instruments of statecraft.

The concept is derived from the Hanna Barbera cartoon universe: the Flintstones’ Bedrock was a prosperous, comfortable, and low-tech society. By contrast, the Jetsons’ Orbit City was where the productive future had already arrived. The choice ahead for Americans is not so cartoonish, but the geometry is the same: the United States must rebuild world competitive manufacturing, or it will live in Bedrock while China and its trading partners race toward Orbit City. In this scenario, America would remain wealthy but stagnant, surrounded by neighbors moving past it.

Consider that when China restricted rare earth exports in April 2025, the United States backed down for a simple reason: it had no leverage. The Pentagon’s dependence on Chinese rare earths is a serious, long acknowledged problem, and it showed. When China offers an ally a discounted 5G network, Washington responds with a diplomatic letter.

The commercial picture is no better. China is now Australia’s top source of new vehicles. Canada, which imposed a 100 percent tariff on Chinese cars in 2024, has since walked it back to 6.1 percent for forty-nine thousand Chinese EVs and is now weighing whether to host Chinese car plants. Britain sits at 10 percent. New Zealand charges nothing.

When America’s closest allies, its Five Eyes partners, open the door to Chinese vehicles, American economic statecraft is at risk. Without competitive manufacturing, economic statecraft is a bluff.4 China has spent thirty years preparing to call it. The survey that follows contrasts America’s Bedrock stasis with Orbit City’s dynamism, further highlighting what’s at stake.

Bedrock and Orbit City

Fred Flintstone was content in Bedrock, unaware of the possibilities in Orbit City. To understand how supply chain mastery unleashes and democratizes innovation, consider what China, which has become a virtual country of Orbit Cities, is actually producing.

In industry, the scale is foundational. China’s investment across every stage of the electric power system has shifted global market trajectories in favor of Chinese manufacturers; in non-fossil energy, what the West calls “clean energy” China treats as strategic: derisking the Malacca Strait, balancing coal dependence, and building out the electric stack. The result, per the Oxford Institute for Energy Studies, is a coevolved ecosystem where policy, entrepreneurship, manufacturers, and customers have developed together rather than in sequence.

In commerce, the drone sector alone numbers seventy-seven thousand companies. Applications span agricultural inspection, planting, pest control, and fish feeding; robot police dogs handle security patrol, crime scene investigation, and covert reconnaissance; traffic drones and humanoid traffic directors manage roads; and robots come in every shape, mobility type, and form factor imaginable. Smart logistics parks integrate hardware and software to manage people, vehicles, goods, and warehouses at scale; where Amazon has built one proprietary system, China’s open supply chain supports dozens of competing vendors. Stranger niches are filled too: surveillance systems that pay for themselves; self-service dog washing machines; and even a robot wolf, originally a Japanese design, built to frighten away wild animals.

For consumers, the product range is similarly broad: baby gear across dozens of categories; pet products including drying pods, weight-sensing fountains, self-cleaning litter boxes, and wireless vacuums; smart washing machines; and elder care robots addressing general assistance, incontinence and bedridden care, heavy lifting for nursing staff, and home care. Autonomous school transport is already operating. On humanoid robots specifically—an area where U.S. firms are competitive—China is already running them in public races, publicity stunt or not.

The case for urgently rebuilding American manufacturing, securing critical supply chains, and stimulating domestic innovation is not seriously contested. What is at stake is not only economic and strategic resilience, but the question of who defines what the next generation of everyday products looks like and who makes them.

Yet another example of the disconnect between Bedrock and Orbit could be seen when Ford CEO Jim Farley visited China in May 2024 and returned a changed man. Despite over two decades of doing business there, what he found in China’s automotive landscape apparently caught him off guard. His verdict was unambiguous: “This is an existential threat.”

The supply chain logic behind that threat is structural. According to the U.S.-China Economic and Security Review Commission, advances in one sector rapidly trigger breakthroughs in adjacent ones, giving China a compounding advantage in successive cycles of iteration and innovation. This is reinforced institutionally: China has built systematic links between researchers and companies specifically to accelerate commercialization. And when Rhodium Group mapped these supply chains at the six-digit classification level, the resulting picture made the depth of China’s supply chain dominance hard to dismiss (see notes for links to graphics).5

The lesson of this data is twofold. First, China has done a tremendous job over a long period of building supply chains across its economy. Second, those supply chains are deeply interrelated: America cannot just support a single supply chain layer (such as rare earths), a single company (such as Intel), or a single industry (such as EVs or semiconductors) and expect that the full range of supply chains upstream and downstream of those selected areas will appear of their own accord. They will not.

One prominent example of the Bedrock trap lives in the technology stack that will run the next economy, 6G telecom infrastructure. Lucent Technologies, America’s last significant domestic mobile network infrastructure company, was acquired by Alcatel in 2006. The United States now depends on Nokia, Ericsson, and Samsung for equipment carrying military communications, commercial data, and industrial control systems. China fields Huawei, which has 31.3 percent of the global RAN market share, $118.6 billion in annual revenue, and $25 billion in R&D (more than Nokia and Ericsson combined), serving 169 countries, with dominance in 160 and sole-source in eighty-three. When the United States asks an ally to exclude Huawei, it asks that ally to pay a premium for a product with a smaller installed base and fewer engineers per market, while offering only security arguments and diplomatic pressure in return.6 As Robert Atkinson documented in these pages in 2020, the loss of Lucent was a foreseen and foreseeable policy failure; it is now structural.

6G compounds the stakes. The standard will integrate AI inference natively into the network, delivered via edge architectures with sub-millisecond latencies, enabling autonomous factories, vehicles, and robots.7 The country that controls the network has a preferential advantage in deploying AI on it, governing latency and throughput by application. The implication is that, regardless of price, efficient AI delivery depends on the network placing control and preference for the AI stack in China’s hands for most of the world.

China has mobilized to lead 6G, with ITU candidate technology submissions due mid-2027 and standards finalization set for 2029–2030. The window to shape 6G architecture, through patent positioning and standards contributions, is effectively closing soon.

5G/6G is, in this sense, a persistent strategic dependency of the same character as the rare earth processing capacity the administration is now rebuilding through the Office of Strategic Capital (OSC) and the Defense Production Act, and the semiconductor fabrication capacity that the chips and Science Act seeks to underwrite. All three are layers of the productive base that a country cannot do without if it intends to practice economic statecraft or lead in future technologies. All three were allowed to atrophy under the assumption that markets would supply them on demand. All three now require deliberate government facilitation to restore.

5G/6G, however, presents a different solution geometry from the other two. The legislative package proposed in this essay is a framework for rebuilding broad-based manufacturing through free market incentives. It will not, on any timeline that matters, produce a U.S. RAN player capable of competing against Huawei. Neither Ericsson nor Nokia, in their current form, has the balance sheet, scope, leadership, or scale to mount that competition on its own. Each carries an enterprise value of roughly $37 to $68 billion, compared with Huawei’s $28 billion in 2025 on R&D alone. Nokia may be seeking an Nvidia rescue in an AI-RAN pivot; SpaceX has acquired EchoStar’s spectrum assets. Neither move closes the gap.

What the moment requires is a corporate, not a legislative, action: facilitated U.S. majority ownership of Ericsson, Nokia, or both, ideally through acquisition by a large U.S. technology company with the balance sheet and worldwide footprint to absorb and revitalize the asset. Cisco, Apple, Amazon, and SpaceX are among the candidates; a U.S.-led consortium with Treasury, the Development Finance Corporation, and sovereign wealth cofinancing is a viable fallback structure. The merged entity needs founder-driven governance, aggressive long-term R&D commitment, and melding into U.S. economic statecraft tools across grant, loan, subsidy, trade, and tariff programs.8 The 2027 vendor selection window is real, narrow, and closing.

We name the 5G/6G case here because it is a clear demonstration of the Bedrock trap, and because it is the most consequential single action the administration could take in the manufacturing statecraft domain that the broader legislative package does not by itself address. It sits parallel to the package proposed in this essay: a related, complementary, and time-critical imperative.

The 5G/6G stack is a concentrated dependency, with the United States having effectively no domestic players. Vehicles are the opposite case in nearly every dimension: large, distributed, deeply American, and yet losing ground every day. Farley’s “existential threat” verdict, quoted above, captured a specific reality. The auto sector is the most consequential single battleground in the broader Orbit City competition. U.S. exports as a share of automotive output have fallen steadily since 2010 from 20 percent to about 12 percent. Indeed, looking at the Australian market, highlighted earlier, should scare every non-Chinese car manufacturer.

No other manufacturing sector compares with the automotive industry’s economic weight. The U.S. auto industry drives $1.2 trillion into the American economy each year, or 4.8 percent of GDP. It supports approximately 10.1 million jobs and was responsible for $730 billion in paychecks in 2022. Every dollar of vehicle manufacturing generates an additional $4.23 in economic value through suppliers, dealers, financiers, and aftermarket services. The automotive sector is the single largest pillar of the American industrial base, accounting for close to 30 percent of U.S. manufacturing value-added. The automobile market, at $4.4 trillion, is seven times larger than the mobile device market.

Its reach extends across the economy. The auto supply chain consumes semiconductors, batteries, motors, sensors, steel, aluminum, magnets, plastics, glass, electronics, and software in volumes that support the demand signal for every layer beneath it. It is a key demand driver of semiconductors outside the data center, one of the largest single industries employing operator-class engineers, machinists, and technicians, and, outside the mobile phone/device, the principal civilian platform on which the technologies of the next economy are being integrated: high-density battery storage, AI-driven sensing and autonomy, embedded silicon and software, and edge-computed control systems.

Economic weight, however, is not the same as competitive output, and the gap is now stark. The American auto industry has deployed enormous capital with poor returns; Ford lost money in three of the past seven years. Ford alone spends roughly $8 billion annually on R&D; GM and Stellantis sit in the same range.9 The industry has assembled a deep ecosystem and spent enormous sums; what it has not done is produce competitive products at competitive prices.

The United States is losing this battle badly. China produced 31.2 million vehicles in 2024, nearly three times the U.S. total of 10.5 million, and exported 5.9 million units, with export growth running close to 20 percent year over year. On automation, the gap is equally stark: Chinese OEMs added fifty-seven thousand robots in 2024 versus fourteen thousand in the U.S., giving China roughly 50 percent more robots per car produced and twice the robots-per-worker ratio. In EVs specifically, a Rhodium Group comparison of BYD and Tesla finds a structural cost advantage of roughly $4,700 per vehicle in China’s favor, with direct subsidies accounting for only about five percent of that gap.10 It is difficult to overstate how far behind America’s vehicles are in terms of technology, features, and price compared with China’s.

The autonomous driving picture is worse. German analysts find that China’s share of global innovation strength in driver assistance and autonomous driving rose to 70 percent in 2024; the United States holds 12 percent, and Germany 14 percent. As recently as 2016, Germany held 60 percent. In February 2026, BMW canceled its Level 3 Personal Pilot system and reverted to Level 2, with Mercedes discontinuing its Drive Pilot in the same window. Both have since taken minority stakes in Chinese autonomous driving companies (Mercedes in Qianli, BMW with Momenta, and Volkswagen with Horizon Robotics) in an explicit effort to license rather than build the next-generation stack. The German playbook, documented by MERICS in its report entitled The Bumpy Road Ahead in China for Germany’s Carmakers, is to relocate R&D to China and source from Chinese suppliers to remain competitive at home. It is rational at the firm level and disastrous at the national one: it transfers the technology stack to the competitor in exchange for short-term margin retention.

American OEMs are following parts of the same script. Ford and Hyundai have pivoted to export-led strategies from China, effectively conceding the domestic Chinese market. Stellantis has ceded China to its joint-venture partner. GM has lost roughly half its China sales since 2017. Tesla remains an outlier on costs while relying on a China-based supply chain for cells and components, and it alone cannot anchor a national vehicle strategy. Tesla has also benefited from a long and large subsidy train.

Vehicles are a critical part of the discussion here, alongside 5G/6G, because they are the largest single demonstration of the Bedrock trap in the U.S. economy and because they show, more clearly than any other case, why the package proposed in this essay is necessary but insufficient on its own.

Torsion in the Supply Chain

Another demonstration of the Bedrock trap arises in the industries just discussed: rare earths, semiconductors, and telecom infrastructure. They are effectively “layers” of the broader manufacturing supply chain. Rare earths are at the lowest tier, semiconductor chips are toward the bottom, and telecom equipment is at the top. In between these layers sit multiple upstream and downstream companies supplying the components, circuit boards, wafers, chemicals, production equipment, and hundreds of inputs that make up the final telecom equipment.

When the government prefers one layer over another, that effective subsidy creates supply (via subsidy) that must be consumed by the downstream supply chain. Bringing some of Ericsson and Nokia’s telecom equipment manufacturing to the United States intuitively sounds like a great idea, but its upstream supply chain, components, circuit boards, substrates, and so on do not exist in America. They mostly come from East Asian suppliers, pointing out the need for continued integration with allies and partners. Without incentivizing them or assuring supply, national security is still compromised, chokepoints remain intact, and economic statecraft remains unsure.

The rare earth supply chain presents similar downstream torsion. Military suppliers directly consume magnets and other rare earths. Building a rare earth supply does not imply a domestic commercial drone industry or a cost-competitive battery, component, and manufacturing supply chain for the electric stack, data centers, peak-shaving power storage, or other downstream users of rare earths. Semiconductors, incentivized by the chips and Science Act, have similar downstream supply chain torsion. Packaging, components, circuit boards, assembly and test, and subsystem and system manufacturing occur mostly outside America. For semiconductors to support national security goals, the downstream manufacturing supply chain needs reshoring and rebuilding. These examples reveal the challenge of supporting specific industries and supply chain layers (which, indeed, require specific and exigent support) without supporting their upstream and downstream layers.

The vehicle case discussed above sits atop all of them. Cars consume rare earths in their motors and batteries, semiconductors in their compute and sensors, and telecom equipment in their connectivity. A reshored auto industry whose suppliers depend on Chinese inputs or Chinese-owned companies does not resolve the dependency; it only moves the visible end of it. The proposed legislative package’s broad, technology-neutral incentive structure is designed for exactly this geometry: tax credits that scale with domestic value added at every layer, capital access that reaches the midmarket suppliers in between, and succession relief that keeps the operator-led shops alive long enough to be modernized.

Innovation without production effectively becomes deindustrialization, a gift to whoever can manufacture what America invents. Historically, America invents, China manufactures, and manufacturing is where employment, process knowledge, export leverage, and future innovation capacity accumulate. Alliances without competitive products become a burden, with partners asked to accept inferior goods at higher prices as the cost of a security guarantee. They will defect when the economics, features, and price alternatives are compelling.

Artificial intelligence and advanced robotics, as is well known, will reduce manufacturing employment globally. That raises the stakes for keeping manufacturing rather than abandoning it: when automation compresses labor-cost differentials, the main historical reason to offshore manufacturing disappears. A robot costs roughly the same in Ohio as in Shenzhen. What remains is proximity to customers, IP security, supply chain resilience, defense-industrial base depth, and the requirements of economic statecraft. Frame the real choice correctly: “manufacturing output with AI-augmented jobs,” or neither at all. The familiar “manufacturing jobs versus something better” framing misses what is on offer. If America is not making things, it gets the job losses from automation without the output, the wages, the multipliers, or the export leverage. The communities hit hardest by the China Shock are the same communities where new manufacturing investment now has the largest wage premium and the highest multiplied impact.

China’s two million industrial robots versus America’s 400,000 is a learning-curve story rather than just a labor-cost story. China leads the automation learning curve that will define competitive manufacturing for the next generation. Every quarter of that learning compounds. Embracing the transition, with humans doing set-up, integration, maintenance, and oversight of automated systems while AI optimizes the factory floor and advanced robotics handles the repetitive and dangerous work, is how America gets to Orbit City. Refusing it, or pretending the old job counts return, is how the country stays in Bedrock.

What American Manufacturers Face

Behind the strategic stakes and the supply-chain torsion sits a more prosaic reality: a typical American manufacturer faces three reinforcing constraints, each of which would be sufficient on its own to hinder investment, and each of which compounds the others. Together, they explain why a country with deep capital markets and unmatched innovation capacity is steadily losing the productive base on which its national security, GDP growth, and statecraft depend. They are the analytical spine of the legislative response. Any policy that does not address all three will fail.

(1) The first reinforcing constraint is profitability. And the numbers tell a hard story. Using Aswath Damodaran’s economic value-added dataset, a standard measure of whether a sector earns more than what it costs to fund it, U.S. manufacturers as a whole come out modestly ahead, returning about 8 percentage points more than their cost of capital. That sounds acceptable until you compare it to software (+20 points), business and consumer services (+18 points), or even pharmaceuticals (+9 points).11

It gets worse when you zoom in on the industries that policymakers most want to revive: semiconductors, heavy machinery, primary metals, autos, and auto parts. In these capital-intensive sectors, the spread actually turns negative: firms are, on average, destroying about 2.6 percentage points of value for every dollar invested.12 The industries most central to any serious industrial rebuilding agenda are the least attractive places to put money.

The gap with software is stark. Software earns nearly double the return of broad manufacturing, and roughly five and a half times the return of capital-intensive manufacturing specifically. Capital, reliably and almost mechanically, flows toward higher returns. Over thirty years, that gap has shaped the entire structure of the American economy, and it won’t be reversed by willpower alone.

(2) The second reinforcing constraint is capital scarcity and cost; this problem with the financial system has structural roots, none of which resolve with the business cycle. One structural root is an asset-liability duration mismatch. U.S. commercial banks fund themselves with short-term deposits and cannot profitably hold the seven-to-fifteen-year fixed-rate loans that manufacturing equipment requires. This is the same problem Fannie Mae was chartered to solve for residential mortgages: no deposit-funded bank can hold a thirty-year mortgage against a one-year liability at scale. It has never been solved for manufacturing above the Small Business Administration (SBA) ceiling. Yet another structural root is the missing middle size band. SBA’s combined 7(a) and 504 maximum is now $10 million13; OSC’s effective floor is over $100 million; syndication minimums begin at $100 million; private credit funds prefer PE-sponsored deals with eight-to-twelve-week diligence cycles. The $10–$250 million revenue band, where roughly forty-five thousand midmarket private manufacturers live, is structurally undersupplied.

The third structural root is cost. Middle market U.S. manufacturers borrow at 7 to 9.5 percent from commercial banks and 7 to 10 percent from community lenders and SBA, against Germany’s KfW at roughly 2.76 to 4 percent for SMEs, and China’s state-bank lending at roughly 3.0 percent. A manufacturer evaluating a domestic capex commitment against the same project located abroad faces a 300-to-700 basis-point cost-of-capital handicap before any other factor enters the calculation. None of this is cyclical. The Senior Loan Officer Opinion Survey shows banks tightening C&I lending standards for  twelve of thirteen quarters for large and middle market firms between Q1 2023 and Q1 2026, and for thirteen of thirteen quarters for small firms, with bank summary text repeatedly identifying “decreased customer investment in plant or equipment” as a tightening driver.14 The structural gap persisted through the most accommodative monetary environment in postwar history (2010–2019); cyclical tightening since 2022 has aggravated but not created it.

The deep tech end of the same problem is well documented: deep-tech companies can raise $20 million to prove a scientific breakthrough but struggle to raise the $50 to $300 million needed to build first-of-a-kind commercial facilities. This is where hardware-intensive innovation dies in America. The utilization consequence on the factory floor is direct. A typical U.S. factory runs roughly two thousand hours per year out of an available 8,600; Chinese factories often run six thousand to seven thousand. With the same building and equipment, one Chinese factory produces three times the output. Labor cost does not explain the gap; the explanation lies in capital utilization.

(3) The third reinforcing constraint is leadership and talent; this is downstream of the first two. Talent follows returns and capital. With much of manufacturing systematically earning below or just above the cost of capital, the operator-class talent that built the American industrial base in the postwar decades has been migrating to higher-return sectors for thirty years. The compounding problem is generational. Project Equity estimates that 60 percent of small and midsized manufacturers (approximately 150,000 businesses employing 3 million workers) have owners at or near retirement age; over 100,000 of these face an actual succession event over the next decade, with irreplaceable process knowledge concentrated in the retiring generation. Without intervention, the modal outcomes are private equity acquisition for asset extraction, sale to foreign buyers, or simple closure when no successor offers.

The new founder problem is the mirror image. A young entrepreneur choosing where to start something today compares after-tax economics across software, finance, services, and manufacturing, and rationally chooses what existing tax and capital structures have already chosen for them. Manufacturing’s leadership-renewal problem is the late-arriving symptom of the first two constraints. A thirty-year migration of capital and talent away from a sector with inadequate returns becomes visible when the founders who built the firms originally start to retire.

These three constraints are mutually reinforcing. Low profitability deters capital; without capital, manufacturers cannot modernize to raise returns; and without competitive returns, the next generation chooses software, finance, or services while the current operator generation ages out without successors. No single fix is sufficient. The package proposed here rests on the premise that all three must be addressed at once: tax credits scaled to domestic value added to repair the income statement; a federally-backed loan guarantee facility to strengthen the balance sheet; and estate, gift, and capital gains relief to keep manufacturing businesses in operator hands through the generational transition.

Before moving on to discussion of the package itself, however, we must first appraise what the current administration has already started, why those efforts cannot reach the missing middle, and why tariffs alone will not close the gap.

What the Administration Has Started and Cannot Finish

The Trump administration and the Department of War (DoW) are executing with genuine resources across the top and bottom of the defense supply chain, alongside the One Big Beautiful Bill Act (obbba) passed last year. These efforts are necessary, urgent, and laudable, but they do not reach the decisive missing middle. Analysis may be focused on three sets of efforts:

(1) OBBBA instituted immediate expensing for investment, research, and development; this provides an effective benefit equal to the time-value of the tax shield on covered investments. This is a meaningful break from depreciation orthodoxy. But it does not address the underlying profitability problem: a tax shield can be worth more than the profit there is to shield for young, growing, and expanding companies.

(2) The Office of Strategic Capital (OSC) has been expanded from a $984 million pilot to $200 billion in loan authority under OBBBA; this is the largest defense industrial finance instrument outside wartime, and the OSC is now recruiting a thirty-person Economic Defense Unit to deploy that capital over three years.15 A 2025 executive order on acquisition advanced commercial-first procurement and cut underperforming programs; a follow-on 2026 order restricted stock buybacks at defense contractors and tied executive pay to delivery. Innovation pipelines, including Replicator, AFWERX, NavalX, and DARPA’s SBPO, are bringing nontraditional vendors into the defense base at scale. Replicator-1 alone drew roughly 75 percent of its suppliers from companies that had never previously held a Department of Defense contract. The “new primes,” namely SpaceX, Anduril, Saronic, Shield AI, are gaining ground, with their share of defense tech contract dollars roughly doubling between FY23 and FY25, from about 0.4 to 0.8 percent of the total.

(3) The critical minerals push. obbba appropriated $7.5 billion in critical minerals financing: $5 billion to the Industrial Base Fund, $2 billion to the National Defense Stockpile, and $500 million in OSC credit subsidy enabling up to $100 billion in critical minerals loans. OSC deployed rapidly, with a $400 million equity investment in MP Materials, a $150 million loan for heavy rare earth separation, a ten -year price floor of $110/kilogram for neodymium and praseodymium, and a $700 million conditional loan to Vulcan Elements and ReElement Technologies. Defense Production Act Title III support reached $550 million in fiscal year 2025. The Pax Silica Declaration embeds these efforts in a broader allied framework that, if it encompasses the full AI supply chain, covers the terrain where the next generation of competition will be decided.

The strategy is coherent, but it is structurally insufficient for the problem. There is much more work to do in terms of reimagining and retooling these institutions, starting with the OSC.

OSC’s Institutional Celling

The institutional form of OSC, which is bespoke Wall Street–style underwriting through a small banker corps, implies an average deal size of $278 million to $556 million and 360 to 720 total transactions over a three-year deployment window.16 Even at the larger transaction count, OSC will execute fewer than one thousand deals over its three-year initial authorization. By design, OSC concentrates on top-tier defense industrial primes and a small number of critical-sector flagships. It cannot reach, and was never built to reach, the tens of thousands of midtier suppliers on which the actual production base depends. OSC cannot reach the 230,000-plus manufacturers below the public market and private equity bar who might be dual-use suppliers, constituting the physical substrate of normal production and surge mobilization.

The new primes face a related ceiling. A doubling of contract share from 0.4 to 0.8 percent is impressive growth, but it still accounts for less than 1 percent of total contract dollars. Even if Anduril, Shield AI, Saronic, SpaceX, and their peers triple their share over the next decade (an aggressive projection), they would still represent a small fraction of Pentagon procurement. The vast bulk of defense production will continue to flow through traditional primes and the supplier base feeding them.

The Defense Innovation Unit’s Replicator initiative is on-ramping nontraditional vendors at an impressive scale, but it missed its drone delivery target, and defense tech awards remain below 0.8 percent of total DoW contracts. The innovation pipeline is genuine, but its translation into production volume is not yet where it needs to be.

The DoW must rightly support its direct requirements, which drive its suppliers toward a combination of vertical integration to meet Buy America standards and exquisite solutions with custom built inputs. Bridging the gap to scale commercial products pioneered from Pentagon demand, the historical diffusion path for defense technology requires something more. Should the DoW insist that vertical capabilities, such as motors, energetics, and other required inputs to its contracted platforms, be spun out as separate companies from these new defense primes? Should these new primes be incentivized or required to resell the capability to other players? What is the mechanism for diffusion? These questions need addressing.

The scale gap in critical minerals remains daunting. China processes 91 percent of the world’s rare earth elements; the United States is 100 percent import-dependent on sixteen critical non-fuel minerals. Building from near zero to meaningful scale is a decade-long project. While administration and DoW actions are to be applauded, Congress should be aware of and monitor possible contingent liabilities. The $500 million in OSC credit subsidy backing $100 billion in critical minerals loans implies a 0.5 percent loss ratio for an industry that barely exists, is not profitable without subsidy, creates significant ecological liabilities, and competes against a Chinese sector that is thirty years ahead.17 The various bespoke deals being struck each have different elements of loans, equity, warrants, crowded-in investors, SPVs, and offtake agreements (some multistage). Congress should be monitoring these to understand how risk and reward flow. The expected loan loss ratio projects a level of confidence that the underlying conditions do not yet support. The same can be said for the $100 billion of OSC loans with a 1 percent projected loss ratio to be allocated to defense requirements.

Govini’s Ark.ai platform identifies roughly forty-five thousand distinct Chinese suppliers embedded in U.S. manufacturing supply chains, up from ten thousand in 2005. Approximately seventy thousand U.S. manufacturing companies import from those suppliers, distributed across America’s 240,000-plus manufacturers. Of these, about 9,300 to fourteen thousand are Tier 1 defense suppliers, of which OSC can plausibly reach a small percentage. The remaining midtier manufacturers with China exposure, and the roughly 170,000 manufacturers with no defense business at all but which are still competing and operating in America, require a different institutional instrument.

They need access to capital at financeable rates. They need profitability to justify long-cycle investment. They need leadership continuity through generational succession. OSC, Replicator, and Pax Silica do none of those things. They were not built to. The mark-to-market moments are accumulating. Chinese firm DJI ships drones across consumer, commercial, and agricultural categories at volumes that dwarf any U.S. production target, including the Pentagon’s emergency program of 340,000 small UAS over four years; the manufacturing depth that supports DJI’s volumes is the depth that a Pentagon emergency program by itself cannot replicate.

The United States reportedly consumed between 14 and 25 percent of its thaad stockpiles over twelve days, supporting Israel against Iran, with median estimates near 20 percent, at production rates that are a fraction of what sustained conflict requires. The mobilization that converted American industry to wartime production in 1942 leveraged corporate leaders who understood their supply chains and could redirect them. Today, we confront what might be called the “Freedom’s Forge Fallacy”: we lack the supply chain, production skills, and raw material base to replicate that mobilization. Modern commercial and defense production is orders of magnitude more complex than it was in 1942. Rebuilding requires broad-based market incentives reaching thousands of manufacturers, not concentrated flagship investments alone. These interdependencies form a web. With the right incentives and transition capital, the web can be rewoven rather than simply cut, much less surgically excised. The legislative package proposed here is the institutional instrument the missing middle requires.

Why Tariffs and Top-Down Programs Cannot Close the Gap

Tariffs provide a price signal. But they do not provide the transition capital, the competitive income statement, or the entrepreneurial incentives required to convert that signal into domestic production capacity. The result is two distinct failure modes, both visible in the post-2018 evidence.

The first is geographic arbitrage. Post-Section 301 trade data confirm the pattern. USITC analysis of 2018–2021 data found that Chinese imports in affected sectors fell 13 percent, while domestic production in those sectors rose only 0.4 percent. Third-country suppliers (Vietnam, Mexico, Cambodia, Taiwan) captured the large majority of diverted demand. Selective tariffs without investment incentives are a guaranteed path to Vietnam’s import substitution, not American capacity. Companies shift final assembly to Southeast Asia while retaining Chinese inputs, components, and tooling, circumventing tariffs without any domestic manufacturing benefit.

The second is rent capture. In industries with concentrated suppliers and limited domestic alternatives, U.S. companies pass some of the tariff onto consumers, pocketing the additional profit. The protection is real; the productive response is not. This is the textbook outcome when tariffs are imposed on industries without simultaneous investment incentives that make capacity expansion the profit-maximizing response. The firm does not invest because investment is risky and the tariff income is certain, and because the tariff regime itself can be reversed by the next administration. The result is consumer-paid rents flowing to existing owners, not new factories. Without an investment incentive that pays out on capacity built, tariffs in concentrated industries function as a subsidy to incumbents rather than a spur to entrants.

Let’s review tariffs against the financial elements required to reshore and rebuild manufacturing: profitability, capital availability and cost, and talent. Tariffs do not improve profitability for export-oriented or domestically competing manufacturers, except that manufacturers rent-seek tariffs by raising prices, raising incumbent margins without inducing new capacity. In general, shift sourcing to low-tariff countries and encourage transshipment, rather than domestic capacity growth in the United States. An estimated $300 billion worth of goods subject to recent tariffs reach the United States from Southeast Asia and Mexico annually, avoiding the levies.

Tariffs cannot provide capital at financeable rates; they change the import price and leave the financing question untouched. Two further problems compound these gaps. Tariff walls shield manufacturers from the competitive pressure that drives world-class cost and quality; the goal is Orbit City, not Bedrock with a moat. And tariff regimes carry their own form of uncertainty: a change in the party in power in 2029 could reverse current policy overnight, so firms making fifteen-year capital commitments now must price that risk in.18 Sectoral tariff programs address one layer of one supply chain at a time, leaving the upstream and downstream torsion already described, and reach only the handful of firms a small banker corps can underwrite directly.

The Limits of Sectoral Programs

Sectoral programs do not close the gap either. The chips and Science Act and similar vertical instruments build capacity at one layer of the supply chain. But they do not incentivize the upstream suppliers of equipment, chemicals, and specialty materials that feed the favored facilities, nor the downstream companies that embed the outputs into final products. chips will build fabs; it will not build the precision-machining shops that make the tooling, the specialty-coatings operations that finish the wafers, or the forgers and component suppliers without whom no fab can operate.

Sectoral import limitations also fail. In December 2025, the FCC blocked new models of foreign drones while grandfathering older ones. That sounds tough, but it means U.S. makers can win at home, to the degree such firms exist, by matching yesterday’s Chinese model. Meanwhile, Beijing speeds ahead to Orbit City, often using American-designed chips produced in Taiwan. Drones are only a single manifestation of the drone, robot, and droid ecosystem within the electric stack that will soon drive large parts of the economy. Blocking imports without supporting the entire drone supply chain to expand capacity and production to create scale and price leaves America once again stranded in Bedrock.

The March 2026 ban on foreign routers did the same, banning new models of foreign-produced routers as they pose “an unacceptable national security risk,” creating both supply chain vulnerability and “a severe cybersecurity risk that could be leveraged to immediately and severely disrupt U.S. critical infrastructure.” Again, the FCC limits imports without a supportive domestic policy for competitors to build. The argument that domestic players will step in rests on a faulty assumption: routers are designed and assembled from mostly purchased chips, boards, components, housings, wire, and so on. The entire supply chain needs support. Three ironies are worth reviewing. First, the ban follows the same pattern as the drone ban: existing routers, with the security risks the FCC is worried about, are exempted. Second, companies can apply for exemptions, which Netgear and Eero (Amazon) have already received, continuing to import security risks in new models and creating the same ironic result: Americans will live in Bedrock while the rest of the world continues to buy updated routers. Third, the U.S. government exempted itself from the ban, leaving consumers and companies in Bedrock City while the government continues to an insecure Orbit City future.

Another significant first-order problem with sectoral and product-level programs is that they are, by definition, narrow, controversial (to pure free market advocates), political (if Congress needs to approve them), and therefore slow, and scattered (some things that need them don’t get them).

One significant example is a subset of personal protective equipment: nitrile gloves, which, along with so many other items, were in short supply during Covid. As detailed by Scott Maier and Garret Murch in this journal, nitrile gloves serve far more than health care, touching on dozens of critical supply areas. Yet, the “United States remains almost entirely dependent on foreign supply for nitrile gloves, with effective control of the market concentrated in China and Chinese-linked supply chains.” Five years post-Covid, why is this still true?

The argument here is not against sectoral bans but for the broader supply chain support that gives them a chance to drive domestic production rather than merely block imports. Neither closes the constraint gap. Tariffs and sectoral programs each address something real, but neither addresses the three constraints identified above. The structural problem is horizontal, distributed across the 240,000 firms that constitute the production base. The legislative package introduced below is the necessary complement: tariffs change the import price, sectoral programs build specific verticals, and the package makes it more attractive to be in the manufacturing business at every layer and every firm size.

The Legislative Package: MISA, Mannie Mac, MERA

The package consists of three bills organized as four tools. The four tools are matched to the aforementioned constraints faced by manufacturers. Together, they fix the structural incentive failure that has defined American industrial policy by default, namely a tax and capital market structure that favors software over steel, services over production, and financial extraction over physical investment.

(1) The Manufacturing and Industrial Security Act (MISA): this act would pair two transferable tax credits tied to domestic value-added (DVA): the Manufacturing Incentives Now Act (MINA, §38A19), which reaches most 240,000-plus naics 31–33 manufacturers20, and the do it now Act (§38B), which adds enhanced credits across fourteen strategic sectors and their supply chains. Both improve the income statement, making domestic production more profitable from startups to multinationals.

(2) The Manufacturing Finance Corporation Act (Mannie Mac): this institution would be a permanent wholly owned federal corporation that guarantees manufacturing loans originated by participating private banks, compresses borrowing costs for midmarket manufacturers from commercial rates of 7 to 10 percent down to 5.7 to 6.2 percent for general manufacturing, and 3.5 to 4.5 percent for the strategic sectors. This brings U.S. financing costs into the range of Chinese state-bank lending on the top priority tiers.

(3) The Manufacturing Entrepreneur Rewards Act (MERA): this program would deliver estate, gift, and capital gains relief to qualified manufacturing businesses held for at least ten years, addressing the leadership-renewal problem posed by an impending wave of retirements in the manufacturing sector. We may now dig deeper into each part of the package.21

Figure 1: Summary of the Three Bills Organized as Four Tools

The Manufacturing and Industrial Security Act establishes two new general business credits under Internal Revenue Code §38A and §38B. MINA (§38A) reaches most naics 31–33 manufacturers.22 It does not pick sectors or technologies. It rewards domestic content at every layer of the supply chain, in proportion to the actual value created in the United States. A manufacturer climbing the value-added stack, say, by moving from final assembly to in-house component fabrication, adds DVA at every rung; the credit structure rewards that climb. Credits stack across the supply chain: each rung of suppliers benefits from counting all the rungs below it, stacking benefits for using domestic manufacturers. For general manufacturing, increasing DVA to 100 percent results in an effective tax rate of 15 percent, compared with the 21 percent statutory rate. Credits are transferable,23 allowing manufacturers with limited tax liability to monetize the credits through sale to investors; elective payment is also available.24 The sunset structure is ten years at full credit followed by a seven-year phase-out, long enough to drive multi-decade capital investment decisions while building fiscal discipline. chips Act tracking data confirm that credit certainty over a decade-plus horizon drives facility investment decisions.

The choice to extend MINA universally across naics 31–33, with only statutory and high-margin exclusions (detailed in footnote 3), is a deliberate policy decision rather than an analytical oversight. The alternative is a winner-picking debate over which sectors merit support and which do not: textiles in or out, food processing in or out, furniture in or out. That debate has no endpoint and no clean analytical resolution because every sector argued for exclusion has a member of Congress arguing for inclusion. The universal scope accepts that the credit will overcompensate for the loss of manufacturing in some nonstrategic sectors. In exchange, it removes the partisan winner-picking dynamic that has stalled every prior attempt at a broad-based manufacturing tax credit, and it preserves the structural argument that what America needs is a competitive manufacturing base rather than a curated list of preferred subsectors. §38B addresses strategic discrimination through enhanced credit for the fourteen priority sectors; §38A addresses the breadth.25

The do it now Act (§38B) establishes an enhanced credit for manufacturers operating in fourteen strategic sectors with national security relevance: semiconductors; rare earth and critical minerals processing; advanced materials; telecommunications equipment; aerospace and defense components; shipbuilding; pharmaceuticals and biomanufacturing; electric vehicles and battery supply chains; industrial machinery; advanced energy systems; and four additional sectors specified in the bill text covering robotics, specialty chemicals, machine tools, and precision optics.26 The enhanced credit stacks with MINA, producing a combined effective rate that can reach 5 percent for strategic sector manufacturers at 100 percent DVA, creating a specific incentive to maximize domestic content where supply chain dependence is most strategically consequential. The bill reaches not only sector primes but also their entire supply chains: a precision-machining shop supplying aerospace components qualifies; a specialty coatings operation serving semiconductor fabs qualifies. The sunset structure is twelve years at full credit, followed by a ten-year phase-out, longer than MINA to reflect larger capital commitments and the investment horizon required to compete with Chinese five-year plans.

Mannie Mac could do for manufacturing what Fannie Mae did for homeownership: create a federally backed risk-sharing mechanism that makes bank lending to an underserved market systematically more available and more affordable, without replacing private underwriting or permanently expanding the government balance sheet. Different from Fannie and Freddie, Mannie Mac is structured as a wholly owned federal corporation rather than a GSE, thereby placing the structure itself outside the implicit guarantee ambiguity that led to the 2008 failure.

The institutional design is deliberately narrow. Mannie Mac would not lend directly, set loan quotas, or operate as a permanent balance sheet institution. It would guarantee manufacturing loans originated by participating private banks, sharing risk through a structure that allows banks to lend at rates they could not otherwise offer while retaining underwriting responsibility. Banks make the loans; Mannie Mac guarantees them; the manufacturing sector gets capital at rates it can afford.27

Mannie Mac serves a company cohort no current program reaches at scale. SBA’s maximum loan is $10 million; OSC’s minimum effective deal is over $100 million. Mannie Mac explicitly targets the $10 million to $250 million band, where about fifty thousand American manufacturers actually live.

The Fannie Mae parallel is instructive as both a functional model and a cautionary tale. Mannie Mac’s structure deliberately diverges from the GSE template that failed in 2008. It is a wholly owned federal corporation, with governance modeled on the Tennessee Valley Authority and operations modeled on Ginnie Mae, rather than a publicly traded enterprise resting on implicit federal backing. The federal commitment is explicit and statutory, codified as a $200 billion Treasury Purchase Agreement in the bill, rather than implicit and contested. Mannie Mac has no shareholders to wipe out and no privatization endpoint to navigate.28

For the strategic sectors specifically, Mannie Mac’s compression of borrowing costs to roughly 3.5 to 4.5 percent brings U.S. financing into the range of Chinese state-bank lending on the top two priority tiers, closing most of the structural cost-of-capital disadvantage that has been a central driver of U.S. manufacturing capital flight.29

The Manufacturing Entrepreneur Rewards Act addresses the third structural constraint. Over 100,000 manufacturing companies are expected to face ownership transition over the next decade as baby boomer founders retire, risking the loss of valuable assets and tacit knowledge. MERA creates the conditions for a different outcome: founder-level entrepreneurs taking over, investing in modernization, and building for the long term. MERA operates along two tracks.

Track One is estate and gift tax relief. Qualified manufacturing businesses held for at least ten years receive relief on gift and estate taxes upon transfer, up to $250 million in relieved value. The ten-year holding requirement rewards genuine long-term ownership rather than financial engineering. Anti-avoidance provisions prevent structured circumvention of qualification requirements. The $250 million cap is calibrated to the practical estate-tax upper bound: a roughly $265 million business generates approximately $100 million in estate tax liability at the 40 percent marginal rate, a ceiling that MERA effectively eliminates. The relieved value eases the transition to the next generation, as the business does not need to either sell to a financial partner or take on large debt to cover taxes, which might be expensive or impossible to obtain.

The program is loosely based on Germany’s small- and medium-sized succession ecosystem. It supports generational transfer through subsidized KfW financing, inheritance and gift tax protections, and national successor-matching platforms such as nexxt-change. Historically, these policies helped preserve the Mittelstand’s long-term industrial continuity.30

Track two is capital gains exclusion. Sellers of long-held qualified manufacturing businesses to outside buyers receive a capital gains exclusion on proceeds, modeled on QSBS treatment but applied specifically to manufacturing businesses.

Together, the two tracks address both sides of the succession transaction. Track One makes family continuity more tax efficient. Track Two rewards entrepreneurs and business operators, not financial investors, who successfully build domestic manufacturing businesses on exit. The combined effect: more of America’s manufacturing base stays in the hands of operators focused on production rather than asset extraction.

MERA addresses seller economics. The buyer-side counterpart is Mannie Mac, which would have a Participating Financial Institution network that could finance operator-buyer acquisitions of family-owned manufacturers at financeable rates that no private credit fund will offer for sub-PE-scale deals. The two tools function as a coordinated succession instrument: MERA on the sell side, Mannie Mac on the buy side. A manufacturing operator who could not previously assemble both the after-tax proceeds the seller requires and the acquisition financing the deal requires can now do both. The package carries a ten-year cost of $1.26 to $1.50 trillion, with a projected return of 1.7 to 3.9 times that investment.31

Three caveats are honest to put forward. First, this is the proponents’ scoring; the Joint Committee on Taxation’s methodology may produce different numbers. Second, the cost is gross before behavioral response; the package’s design assumption is that broad-based, automatic, technology-neutral incentives induce more investment than discretionary allocation does.

The same political-reversibility logic for tariffs applies, in fairness, to MISA and Mannie Mac. The package’s response rests on three structural durability factors. First, modifying a tax credit or a federally chartered corporation requires affirmative congressional action and the president’s signature, a materially higher bar than executive tariff modification. Second, adopting an existing credit transferability regime (IRC §6418) creates a market for MISA credits, already operating at scale for Inflation Reduction Act credits (with an assured U.S. government purchase option at 85 percent of the credit). Third, Mannie Mac’s wholly owned federal corporation status, with and explicit $200 billion Treasury Purchase Agreement and supermajority-protected DVA and fee floors, is a statutory commitment, with TVA’s ninety-year record across many administrations as an empirical example. The package is not immune to reversal; it is materially harder to reverse than executive trade actions.

Anticipating the Strongest Objections

A package of this scale will draw three principal objections, which deserve direct answers.

The first is the free market objection: This is industrial policy by another name. It is, but it is industrial policy that does not pick winners. It is industrial policy that corrects the structural tilt of policy and incentives toward investing away from manufacturing. MISA does not select sectors (beyond §38B’s enumerated strategic list, which corresponds to the existing National Defense Strategy and the OSC priority sectors) and does not select firms within them. The DVA-scaled credit rewards domestic content automatically, in proportion to the value it creates. Mannie Mac does not originate loans; private banks underwrite, with Mannie Mac sharing risk on a fee basis. MERA rewards operators the holding period for manufacturing companies, not the business model. The argument rests on structure, not sympathy: the existing tax and capital market structure, by accident of design rather than intent, systematically underinvests in manufacturing relative to its strategic and economic value. The package shifts that market structure. It is market-based and automatic, not discretionary.

The second is the fiscal objection: We cannot afford another $1.5 trillion on top of OBBBA. The framing of the cost matters. First, this package is about rebuilding national security and future GDP in the era of AI from the ground up. Its goal is to be consistent with the top-down DoW and administration’s efforts at the company and sector levels, and with the supply chain required to turn those efforts into actual manufacturing capacity. Second, the package’s projected return on investment exceeds one-to-one even at the low end of estimates: this is investment, not consumption. Third, the alternative is not zero cost: continued deindustrialization carries its own fiscal price in lost wage tax base, social safety net spending in deindustrialized communities, and the eventual cost of a wartime crash mobilization. Fourth, the cost is gross-of-revenue from increased manufacturing activity; even modest reshoring of currently imported goods generates federal tax revenue that offsets a meaningful share of the headline number. Fifth, the largest line item (MERA) has substantial fiscal pretexts already built into the tax code: QSBS for new C-corporations, §1014 step-up basis for inheritance, and valuation discounts on closely held interests. MERA channels existing tax preferences toward a defined policy goal benefitting operators, not financial investors, rather than creating an entirely new fiscal commitment.

The third is the comparative advantage objection: If China is better at manufacturing, the rational move is to let China make things while America makes software, finance, and services. This is the strongest objection—and the most decisively incorrect one. Comparative advantage as a static doctrine rests on three assumptions that no longer hold: that the trading partner is benign, that the goods outsourced are not strategic, and that peacetime endures. Each fails against the actual China relationship. China builds leverage continually. The outsourced goods now include precisely those required by national security. Peacetime ended some time ago; we are in a cold war, regardless of what trade agreements or commissions are agreed.

The Ricardian framework was developed in 1817 between trading partners with no military rivalry; applied uncritically to an adversarial relationship with the world’s second largest economy, it counsels strategic surrender. Free trade with allies is one thing; structural production dependency on a strategic adversary is another.

The package is built for the institutional vehicles available to it. MISA fits naturally within reconciliation as tax policy under Ways and Means and Senate Finance jurisdiction. Mannie Mac is a Financial Services and Banking Committee bill; the Defense Production Act reauthorization window in September 2026 is a plausible vehicle, given DPA’s existing manufacturing-finance authorities. MERA is also a Ways and Means / Senate Finance vehicle, and its succession relief framing aligns with bipartisan family business interests on both committees. The bills can move as a single MISA package or as separate components; the analytical case holds either way.

The package does not address every component of manufacturing competitiveness. Government-sponsored R&D, permitting and regulatory reform, workforce retraining through expanded apprenticeship programs, and state-level economic zones all matter. Arizona and Ohio’s semiconductor zones demonstrate what focused state competition for manufacturing investment can achieve. These domains require parallel action. But none of them substitutes for the structural fix to the federal incentive architecture that the legislative package provides. It is up to states and localities to supplement federal policy with their own offers of tax, workforce, land, economic zones, and so on.

The Allied Math Only Works If We Can Make Things

The United States has 330 million people, compared to China’s 1.3 billion. We cannot offer a larger market. We must offer a better one. Despite what today appears to be a fraying of traditional allies, the allied coalition (Europe, India, Five Eyes, Japan, Southeast Asia) has the population, capital, technology, and market depth to outcompete China on every dimension. The coalition math works in principle. What it requires in practice is the American manufacturing credibility to make the offer real, and as this essay documents, the credibility is not yet there: Five Eyes partners are buying Chinese cars, German automakers are relocating R&D to China; Nokia and Ericsson are without the balance sheet to fund competitive 6G development. The coalition has the resources. It does not yet have the American offer that catalyzes those resources into a competitive alternative to China.

Pax Silica is an initial institutional architecture. If it encompasses the full AI supply chain, it covers the terrain where the next generation of competition will be decided (and needs to include 5G/6G). But the architecture is only as strong as the products behind it. An alliance with superior standards and inferior products will fail, as NATO’s Huawei exclusion effort has partially failed: allies comply under duress and defect when the economics allow. Australia takes Chinese cars. Canada imports them at a single-digit tariff. Five Eyes partners measure trade-offs and decide. The pattern repeats whenever the U.S. cannot offer a competitive product at a competitive price.

The supply-chain shocks of the past decade (e.g., rare earths in 2010 and 2025, Covid-19 active pharmaceutical ingredients and personal protective equipment, Ukraine drone production requirements) all follow the same pattern: strategic signals sent, years of institutional non-response, and crisis-driven overcorrection at orders of magnitude greater cost. Rare earths are the visible symptom of a deeper systemic failure: the inability to translate economic statecraft authority into capability.

A historical objection deserves a direct answer. The United States exercised substantial economic coercion through the 1990s and 2000s precisely as its manufacturing share declined, using financial instrument statecraft (through SWIFT exclusion, secondary sanctions, entity list designations, dollar-clearing controls) that operates independently of naics 31–33 output. Those instruments still work against weaker adversaries and isolated targets, and they are still essential to the toolkit. They degrade sharply against a peer competitor with alternative payment rails (CIPS and the multi-CBDC mBridge platform), export control reciprocity (the rare-earth controls of April 2025), commercial product leverage allies actively want, and an embodied-AI transition in which the production base itself is the platform. The 1990s sanctions worked because adversaries had no alternative manufacturing statecraft toolkit. China has built one. The financial instrument toolkit alone no longer compensates.

Nadia Schadlow diagnosed the underlying pathology directly: “Devotion to process replaced attention to outcomes.” The 1,400 U.S. government offices with some responsibility for economic statecraft exist, in theory, to execute a U.S. value proposition. The problem is that no one has defined what that proposition is, and the apparatus was not built to deliver it at commercial speed against a competitor operating in an entirely different way. Process without product is bureaucratic theater. The legislative package is what gives the apparatus something concrete to defend.

Right now, America subsidizes software and taxes manufacturing, not by statute but by structure. The proposed bills change that structure. They are market-based, broad-based, and measurably fiscal. They are an investment in the productive base on which every other instrument of American power depends.

Five consequences converge on a single point. You cannot practice economic statecraft without competitive products to offer, withhold, and build. You cannot defend the country without the industrial base to produce what defense requires. You cannot lead the technological future without the physical infrastructure to manufacture it. You cannot make the allied coalition math work without a credible economic offer that closes the gap. And you cannot honor the alliances on which eighty years of American leadership have rested if the goods behind those alliances do not exist at a competitive price and quality.

All roads lead to, or better, lead from, the American industrial base. Bedrock is wealthy and stuck. Orbit City is where the future is made. The choice is in front of us now, and the legislative window is open. Congress should pass these bills.

1 The Houthis’ redirection of world trade away from the Red Sea and Iran’s intermittent control of the Strait of Hormuz, where Chinese ships continued to transit even as others halted, are 9/11-scale wake-up calls we have failed so far to answer.

2 The premise that some allied buyers in some categories will pay a premium for American-origin supply is the demand-side foundation of the statecraft case rather than an empirically established quantity. Three forms of available evidence support the directional claim: allied procurement programs already command preference premia in defense-grade contracts under Buy America, Buy American Act, and Five Eyes frameworks; trusted-vendor RFPs for 5G infrastructure document explicit premia where allied governments select Ericsson or Nokia at a cost above Huawei; and multi-year allied defense procurement contracts (F-35, AUKUS submarine program, Patriot/SM-3) demonstrate sustained willingness to pay U.S.-origin premia where the alternative is China- or Russia-origin supply. The premium varies sharply by category and is small or zero in the commodity sectors where the sidebar’s price-competition concession applies. Quantitative survey work would strengthen the claim. If the world shifts more towards spheres of allies and partners, American leadership must include at least near-competitive product offerings.

3 “Manufacturing” as used in this paper means the full scope of physical production capacity on which national security and economic statecraft depend: (1) naics 31–33 excluding 3121 (Beverage Manufacturing), 3122 (Tobacco Manufacturing), and 3256 (Soap, Cleaning Compound, and Toilet Preparation Manufacturing); (2) energy from ground to point of delivery; (3) critical minerals and rare earths from mine through separation and productization; (4) active pharmaceutical ingredients and biomanufacturing; (5) defense production including munitions and propellants; and (6) the construction of productive capacity itself, including factories, tooling, and industrial plant. The common thread: any physical production process whose absence or adversary control creates strategic vulnerability or forecloses statecraft options.

4 China’s alternative energy efforts, including nuclear, solar, wind, and batteries, manifest in the West as efforts to mitigate climate change. To China, they are a long-term strategy to derisk oil and gas dependencies. We have driven China’s scale and investment in these technologies through market access and subsidies.

5 Camille Boullenois, Malcolm Black, and Alessia Caruso “China’s Next-Generation Industrial Policy,” Rhodium Group, May 11, 2026. In particular, see Figure 31; Rhodium Group’s six-digit supply chain classification chart shows the depth of Chinese supply chain development across categories.

6 In the US, no single program matches the integration of China’s approach, where vendor financing (CDB/CHEXIM), equipment subsidies, and diplomatic pressure operate as a coordinated package. U.S. programs are siloed across FCC, NTIA, EXIM, DFC, and State, with no lead agency and no unified financing vehicle. Domestically, the FCC’s “Rip & Replace” program (Secure and Trusted Communications Networks Act, 2019) funds the removal of Huawei/ZTE equipment by rural carriers and replacement with Ericsson, Nokia, or Samsung gear, but Congress has funded only approximately 40 percent of the estimated need ($4.98 billion requested versus $1.9 billion appropriated), leaving the program chronically underfunded.

See: SCSP, “6G Rollout: Strategy, Spectrum, & Security,” Special Competitive Studies Project (Substack,) September 29, 2025; this source documents the absence of a federal national strategy unifying spectrum policy, standards-setting, R&D investment, and security; the CDB/CHEXIM scale comparison was earlier documented in Carolyn Bartholomew, “China and 5G,” Issues in Science and Technology 36, no. 2 (Winter 2020); this source traced forty-five of forty-seven Huawei projects in Africa to CHEXIM financing and observed that CHEXIM operates ‘far beyond providing export credit to companies in the way the United States’ EXIM Bank does.

7 ITU-R Recommendation M.2160 formally established Artificial Intelligence and Communication as one of six 6G usage scenarios, treating AI compute, model hosting, and inference as native network services rather than overlay applications, a significant architectural departure from 5G. See: International Telecommunication Union, Framework and Overall Objectives of the Future Development of IMT for 2030 and Beyond, Recommendation ITU-R M.2160-0 (Geneva: International Telecommunication Union, November 2023).

8 We address the case in detail elsewhere. See: “The United States Needs a 5G/6G Player,” Rebuilding Manufacturing, accessed June 2026.

9 Domestic OEMs and their battery partners had committed approximately $125 billion across 80-plus EV and battery projects between 2021 and 2024. The end of 2025 brought the reckoning. Ford recorded $19.5 billion in special charges in its Q4 2025 EV strategy reset, including an $8.5 billion writedown of EV assets and the cancellation of the F-150 Lightning; Ford’s EV division had already lost approximately $13 billion since 2023, including $5 billion in 2024. GM recorded roughly $6 billion in Q4 2025 EV-related special charges, including $4.2 billion in cash payments to suppliers whose committed capacity is now unwanted, on top of a $1.6 billion writedown in Q3 2025.

10 The dominant factors are vertical integration (BYD manufactures around 80 percent of Tier 1 components in-house against 37 percent at Tesla and 35 percent at Volkswagen), concentrated and lower-cost R&D (BYD spends $2,302 per vehicle on R&D and administration against $4,021 at Tesla), and supplier payment float (155-day terms at BYD against sixty days at Tesla and forty-three at Volkswagen). Western OEMs have lost roughly 27 percentage points of Chinese market share since 2020, from 62 percent to 35 percent. Volkswagen’s share alone fell from 19 percent in 2019 to 14.5 percent in 2024, overtaken first by BYD and then by Geely. Vertical integration, R&D-per-vehicle, cost-gap, and market-share figures are drawn from: Gregor Williams, “Why Are Chinese EVs So Cheap?,” Rhodium Group, February 19, 2026.

11 Damodaran’s EVA dataset covers publicly listed U.S. firms; the 240,000-manufacturer population targeted by MINA and Mannie Mac is overwhelmingly private mid-market companies. Private mid-market manufacturers face a higher cost of capital (a three hundred to five hundred basis-point credit-spread premium over public investment grade) and lower scale efficiencies than the listed sample. The Tier 1 +8pp and Tier 2 -2.6pp figures are therefore upper bounds on the true private mid-market EVA spread; the actual figures are almost certainly worse.

12 The capital-intensive Tier 2 sample comprises Damodaran’s industry groupings for Semiconductor Capital Equipment, Machinery, Steel, and Auto & Truck, plus Auto Parts, weighted by total invested capital. The negative 2.6 percentage-point figure is the trailing 5-year weighted average; the most recent single-year reading is even more negative. Reproducible computation at: “Why Capital Fled Manufacturing,” Rebuilding American Manufacturing, accessed June 2026.

13 On free-market philosophy: the argument rests on structure, not sympathy. The existing tax and capital market structure, by accident of design rather than intent, systematically underinvests in manufacturing relative to its strategic and economic value. The bills proposed here shift that market structure. They are market-based and automatic, not discretionary. They do not pick winners; they improve the economics of manufacturing as a category.

14 The Q4 2025 SLOOS also documents differentiated bank willingness to lend by firm AI-exposure, with banks more receptive to firms with high AI integration and more cautious toward firms negatively affected by AI disruption. Some portion of the tightening reflects a rational lender response to sector disruption rather than uniform manufacturing discrimination. The “decreased customer investment in plant or equipment” language as a tightening driver appears specifically in the manufacturing-relevant SLOOS releases and predates the AI-exposure tightening signal documented in the Q3–Q4 2025 reports.

15 OSC’s $200 billion loan capacity is funded with a $1.5 billion appropriation, a 0.75 percent appropriation-to-capacity ratio that requires very low-risk loans backed by offtake agreements, long-term contracts, and significant co-investment. That constraint becomes material when compared to Mannie Mac’s manufacturing-empirical loss calibration discussed in Section IV: combined Loss Reserves and Supplemental Loss Reserve of $13.5 billion against $245 to $325 billion in cumulative guaranteed capacity (a 4.2 to 5.5 percent reserves-to-capacity ratio), sized to absorb a 5.25 percent central-case lifetime gross default rate (3.41 percent net loss after 65 percent loss given default).

16 OSC deal-size calculation: thirty-banker Economic Defense Unit scaled to sixty for generosity. At two to four closings per banker per year (consistent with a three-to-six-month average deal timeline typical of bespoke infrastructure lending), the portfolio is 360 to 720 transactions over three years. $200 billion divided by 360 to 720 yields an average deal size of $278 million to $556 million, nearly three times OSC’s largest loan under its original $984 million authority. The 230,000-plus mid-tier manufacturers needing $10 million to $250 million instruments will barely appear in OSC’s deal pipeline, regardless of banker capacity.

17 $500 million in credit subsidy backing $100 billion (half OSC’s $200 billion) in critical minerals loans implies a 0.5 percent appropriation-to-capacity ratio for an industry that barely exists, is not profitable without subsidy, creates significant ecological liability, and competes against a Chinese sector 30 years ahead. Mannie Mac’s combined reserves of $13.5 billion against $245 to $325 billion in guaranteed capacity (a 4.2 to 5.5 percent reserves-to-capacity ratio) make the program between eight and eleven times more conservatively reserved for a much broader universe of established commercial manufacturers, sized to absorb a 5.25 percent central-case lifetime gross default rate and a 12.0 percent worst-case rate under stress overlays.

18 The 2025–2026 empirical record confirms the structural diagnosis. ISM’s Spring 2025 manufacturing forecast projected capex collapsing from +5.2 percent to –1.3 percent as tariff volatility made long-term investment irrational. The actual 2025 outcome recovered to +3.5 percent, with the December 2025 ISM forecast projecting another +3.0 percent for 2026, a recovery, but at a sustained level roughly two percentage points below the pre-tariff trajectory. Manufacturing PMI returned to expansion in January 2026 after twenty-six consecutive months of contraction, remaining in expansion through April 2026, though manufacturing employment continued to contract, with 102,000 manufacturing jobs lost between December 2024 and March 2026, with the majority of those losses occurring after the April 2025 “Liberation Day” tariff announcement. The Joint Economic Committee’s Democratic staff estimated in August 2025 that tariff uncertainty might foreclose $490 billion in manufacturing investment by 2029. Recovery in headline activity confirms rather than refutes the diagnosis: manufacturers absorbed the volatility cost, paid for it in lost jobs and below-trend capex, and proceeded. None of these tariffs affected the items discussed here: profitability, capital, talent.

19 The Internal Revenue Code is the codified federal tax statute. §38A and §38B would be new sections within the general business credit framework of Subtitle A, Chapter 1, Subchapter A, Part IV, Subpart D.+

20 Two categories of naics 31–33 manufacturers are ineligible for the credits. First, naics 3121 (Beverage Manufacturing, including alcoholic and non-alcoholic beverages) and naics 3122 (Tobacco Manufacturing) are permanently excluded by statute. Second, any naics 31–33 industry whose trailing three-year EVA spread (industry ROIC minus WACC) exceeds eighteen percentage points is excluded annually by IRS Notice, unless the industry is also designated a §38B strategic sector. naics 3256 (Soap, Cleaning Compound, and Toilet Preparation Manufacturing) currently meets the EVA-spread threshold and is excluded under this mechanism. In aggregate, excluded industries represent approximately 4 to 5 percent of naics 31–33 value added, reducing the eligible DVA cost base by roughly $100 billion.

21 For MINA, Section 38 of the tax code establishes tax credits. See: 26 U.S.C. § 38, “General Business Credit,” Legal Information Institute, Cornell Law School, accessed June 2026. The DVA scaling mechanism, where Domestic Value Added is expressed as a share of Domestic Cost to Total Costs, is the unifying metric across MISA’s two credits. Credits scale automatically from a floor for minimal domestic content to a 100 percent ceiling for fully domestic production, without bureaucratic allocation or discretionary review. Improving the income statement to improve price competitiveness and returns is the only durable way to incentivize rebuilding manufacturing.

22 Two categories of naics 31–33 manufacturers are ineligible for the §38A credit. First, naics 3121 (Beverage Manufacturing, including alcoholic and non-alcoholic beverages) and naics 3122 (Tobacco Manufacturing) are permanently excluded by statute; these are “brand-based” industries, not national security related, and generally have high margins. Second, any naics 31–33 industry whose trailing three-year EVA spread (industry ROIC minus WACC) exceeds eighteen percentage points is excluded annually by IRS Notice, unless the industry is also designated a §38B strategic sector. In aggregate, excluded industries represent approximately 4 to 5 percent of naics 31–33 value added, reducing the eligible DVA cost base by roughly $100 billion.

23 Under IRC §6418, established by the Inflation Reduction Act and now operating in a mature credit-transfer market. See: 26 U.S.C. § 6418, “Transfer of Certain Credits,” Legal Information Institute, Cornell Law School.

24 DVA implementation requires a Treasury and IRS rulemaking framework, including a DVA reporting standard (building on the cost-accounting standards already used in federal contracting under 48 C.F.R. Part 30), a bright-line anti-manipulation rule for related-party transactions and transfer pricing under IRC §482, and an audit infrastructure leveraging the existing §6418 transferability framework now operating at scale for Inflation Reduction Act credits. The §38A bill text grants Treasury rulemaking authority and includes graduated enforcement consequences under §38A(o); compliance architecture will be developed during the implementation phase with Notice-and-Comment rulemaking before the credit first applies.

25 The geographic-arbitrage and rent-capture failure modes that limit tariffs as standalone instruments apply with reduced force to DVA credits. Geographic arbitrage requires transfer-pricing manipulation across related parties, which is already policed by the IRC §482 audit framework and further tightened by the DVA-specific reporting standard described in footnote 18. Rent capture is partially structural: DVA credits scale with the level of domestic content rather than the increment, by design, because increment-only credits create cliff effects, disincentivize firms already at high DVA from maintaining it, and reward the gaming of base years. Level-based credits reward domestic content at every starting position, recognizing that some firms with stable, high-DVA production will harvest credit from existing operations as the cost of incentivizing the broader population to climb.

26 The §38B strategic sectors are specified by naics code in the bill text to provide legal certainty and prevent administrative scope creep. The fourteen-sector list reflects the May 2023 National Defense Strategy Technology Stack designation, expanded to include four additional sectors with upstream supply chain relevance: industrial robotics; specialty chemicals; machine tools and precision metalworking; and precision optics.

27 The financial architecture: approximately $21.4 billion in federal investment ($18.9 billion main program plus $2.5 billion Patient Capital sub-program under §17, which serves manufacturers in the $200 million to $1 billion revenue band with longer-tenor strategic-sector commitments) supports $245 to $325 billion in cumulative guaranteed loan capacity over the twelve-year initial authorization, at a net twelve-year federal fiscal cost of approximately $8.6 to $9.0 billion. The result is manufacturer access to capital at rates of 5.7 to 6.2 percent for general manufacturing and 3.5 to 4.5 percent for strategic sectors, against U.S. commercial rates of 6.5 to 11.7 percent that currently make long-term manufacturing investment economically irrational for most private companies.

28 For summary details, see: “Mannie Mac—Brief (4-Page, v2),” Rebuilding American Manufacturing, accessed June 2026.

29 Chinese policy bank rates to favored sectors include below-market subsidies embedded in policy-bank structures that are difficult to audit independently. The People’s Bank of China one-year loan prime rate stood near 3.0 to 3.1 percent through 2025, with preferential rates for EVs, semiconductors, and rare earths running materially below that. The directional differential against U.S. commercial lending of four hundred to seven hundred basis points is consistent across the published loan prime rate, World Bank country lending statistics, and independent Western think-tank analyses, even where the precise preferential rate for any given sector cannot be verified.

30 Note that demographic aging, weak successor pipelines, rising regulatory burdens, and modernization costs are now producing a growing succession crisis in Germany, with record numbers of firms considering closure rather than transfer. On the theory that, properly incentivized, either the next generation of a business owner or outside buyers will find American companies worth saving, supporting these companies is a worthwhile endeavor. The free market will weed out the ones that should close.

31 Drawing on BEA input-output multipliers (2.5 times), University of Maryland Inforum model estimates (2.9 times), Mertens and Ravn (2013) production-targeted incentive multipliers (3.0 times), and Bureau of Labor Statistics wage premium data, ranges from 1.38 to 2.95 times the initial expenditure. Even the low end is favorable; the high end is transformational. NAM data show the manufacturing wage premium at 23 percent above the private-sector average ($43.07 versus $35.01 per hour), rising to 27 percent with benefits included. The Economic Policy Institute finds production workers earn 13 to 18 percent more than similarly educated service-industry workers, controlling for age, sex, race, and geography. Reshoring Initiative data confirm that 3.5 to five supply-chain and service jobs are generated for every direct manufacturing position.