global geopolitics

Decoding Power. Defying Narratives.


America’s Biggest Industrial Bet Is Going Wrong

The AI boom is the largest wager in American corporate history, and the three forces working against it – market concentration, Chinese scale and a cracking foreign policy consensus are not being addressed

Opinion Analysis | November 2026

The AI boom is the largest wager in American corporate history, and the three forces working against it – market concentration, Chinese scale and a cracking foreign policy consensus  are not being addressed

Opinion Analysis | November 2026

The American economy has committed itself to what bank research desks and corporate filings now describe as the largest single industrial wager in the history of American capitalism, exceeding the railroads, the electrification of the economy and the atomic energy programme and the wager is not proceeding according to the expectations of those who placed it. This article examines three connected developments that together constitute a structural crisis. The first is the financial architecture of the artificial intelligence programme, in which an extraordinarily concentrated market conceals a wide dispersion of corporate performance and in which positioned capital now expects a correction. The second is the Chinese industrial advantage in economies of scale and scope, which the programme is intended to overcome and which the available evidence suggests cannot be overcome on the terms currently pursued. The third is the cracking of the bipartisan foreign policy consensus that governed American conduct for seven decades, first over Europe and then more sharply over China. The removal of a prominent technology executive in November 2026 is analysed as narrative management of a deteriorating strategic position rather than as a security event, and the distributional consequences of the programme are assessed against the populations that bear them.

The scale of the commitment is the appropriate starting point, because the magnitude of the wager determines the consequences of its failure. Richard Wolff characterises the investment in artificial intelligence as the largest single bet on an industry in the history of American capitalism, exceeding the railroads, the electrification of the economy and the atomic energy programme that were previously the leading candidates for that designation (Wolff, 2026). The characterisation is difficult to verify with precision because the comparison depends on the metric chosen and the period over which it is measured, and the direction of the claim is supported by the observable data on capital expenditure. The major cloud providers have guided to combined capital spending in excess of three hundred billion dollars for the coming year, with the largest share directed toward artificial intelligence infrastructure, and the Bank for International Settlements has noted that the concentration of this spending among a small number of firms is without precedent in the post-war period (Bank for International Settlements, 2025). The commitment is therefore not a speculative projection but a documented allocation of resources on a scale that constrains the alternatives available to the firms that have made it.

The financial structure that has grown around the commitment exhibits the characteristics that historically precede a correction, and the characteristics are visible in public market data rather than in inference. The broad American equity market appears healthy when measured by its headline index, and the appearance is produced by a very small number of companies whose performance conceals the difficulties of the majority. The International Monetary Fund has documented the widening dispersion of corporate performance in its assessments of financial stability, noting that the gap between the most and least profitable firms in the index has reached levels associated with previous periods of market stress (International Monetary Fund, 2025). An active market in short positions against the largest technology firms has developed, and the participants in that market are not hedging against a downturn so much as positioning for one. The distinction matters because a hedged investor wishes to be protected from a fall, whereas a positioned investor wishes the fall to occur, and analysts who have examined the depreciation schedules applied to AI infrastructure, including Michael Burry, have argued that reported earnings overstate the true returns by assuming useful lives for equipment that advances too rapidly to sustain them (Burry, 2025).

The political consequences of the programme have begun to appear in electoral behaviour, and the appearance is the most direct evidence that the consent on which the programme depends is not secure. More than half the states held ballot propositions in November 2026 that were hostile either to the semiconductor industry or to the electricity and gas companies whose output the data centres require, and the propositions passed by overwhelming margins with no recorded instance of a contrary result (Wolff, 2026). Candidates for Congress campaigned on commitments to scrutinise the programme, to slow its expansion and to examine its costs, and a substantial number of them were elected on those commitments. The electoral signal is significant because it demonstrates that the public has formed a judgement about the programme independently of the discourse that surrounds it, and because it establishes that the democratic process can impose constraints even when the interests promoting the programme are extraordinarily well resourced. Martin Gilens and Benjamin Page demonstrated some years ago that American policy responds to the preferences of economic elites rather than to those of average citizens on the great majority of issues they examined, and the AI programme is an instance of the general pattern in which elite preferences prevail until public costs become too visible to ignore (Gilens and Page, 2014).

The Chinese industrial position that the programme is intended to counter rests on advantages that are structural rather than contingent, and the structural character of the advantages is what makes the current strategy appear inadequate. Modern industrial production requires an immense fixed capital investment in long-lived equipment, and the cost of that equipment must be recovered across the units produced, which means that a producer with larger output can charge a lower unit price than a competitor with smaller output. Alice Amsden documented how late-industrialising states in East Asia captured market share by combining scale with state-directed investment and learning-by-doing, and her account of Korea’s steel and shipbuilding industries describes a process that China has repeated at a larger scale (Amsden, 1989). Robert Wade’s analysis of Taiwan’s industrial policy made a related argument about the role of state direction in creating scale advantages that markets alone would not have produced (Wade, 1990). China operates an industrial economy with approximately 1.4 billion people against an American industrial economy with approximately 340 million, and the ratio of roughly four to one provides a domestic market of a scale that no American producer can match.

The advantage of scope compounds the advantage of scale, and the compounding is the part of the analysis that is most often overlooked in American commentary. David Teece established the theoretical basis for economies of scope in a paper that demonstrated how the sharing of inputs across product lines produces cost advantages that firms operating in a single product cannot match, and his framework explains why a firm that has mastered one technology can enter adjacent markets at a fraction of the cost that a new entrant would face (Teece, 1980). The sensors, software and computational architecture required for autonomous vehicles are substantially the same as those required for industrial robotics, which means that a firm that has mastered one domain can enter the adjacent domain at far lower cost. The same logic applies across search, logistics, payments and platform services, and the boundaries between previously distinct industries dissolve as a consequence of the shared technological foundation. Peter Nolan’s research on the growth of large Chinese firms documents how they have used the domestic market as a platform from which to expand into adjacent sectors and then into export markets, and the process he describes is the empirical counterpart of Teece’s theoretical framework (Nolan, 2001).

The removal of Jensen Huang from the chief executive position at NVIDIA in November 2026 belongs within this analysis as an instance of narrative management rather than as an isolated corporate event. The allegations that precipitated the departure concerned a personal relationship and an unsubstantiated claim of foreign espionage, and no American agency has charged the individual concerned with any offence (Wolff, 2026). The absence of a charge is significant because espionage is a serious offence under American law, and the initiation of proceedings would have produced a defence, a hearing and a documented record, and none of those appeared. The structure of the episode therefore resembles a legal proceeding without its substance, and the function of such a structure is to produce a resignation without an adjudication. Daniel Byman and Kenneth Pollack documented a comparable pattern in the management of intelligence failures, in which the attribution of responsibility to a single individual or a single service protects the broader institutional structure from scrutiny (Byman and Pollack, 2007). The narrative that the episode generated attributes the departure to betrayal rather than to the structural pressures that the company faces, and the attribution protects the political system from the recognition that the pressures are material rather than personal.

The industrial policy that the American state has pursued in response to the Chinese position has been substantial and has not yet produced the intended result, and the gap between the scale of the policy and the scale of its effect is itself an analytical finding. The CHIPS and Science Act allocated approximately fifty-two billion dollars to semiconductor research and manufacturing, and the programme has funded fabrication facilities that will begin production over the coming years (Congressional Research Service, 2026). Barry Bluestone and Bennett Harrison documented the process of deindustrialisation that preceded the current policy in their classic study of capital flight, and their analysis of how financial incentives encouraged firms to relocate production explains why the industrial base that the subsidies now seek to rebuild was dismantled in the first place (Bluestone and Harrison, 1982). Seymour Melman’s work on the permanent war economy made a related argument about the diversion of resources from productive capacity to military expenditure, and his analysis bears on the current discussion of how the AI programme is being justified partly by military applications (Melman, 1974). The subsidies address the cost of capital investment, and they do not address the structural advantages of scale and scope that the Chinese position provides, and the difference between addressing a cost and addressing a structure is the difference between a policy that can succeed and one that cannot.

The comparison with the European position is instructive because it isolates the variable that the American analysis tends to obscure. The European Union contains approximately 450 million people, which is a population comparable to the American population and substantially smaller than the Chinese population, and the European market is divided into twenty-seven regulatory jurisdictions whose competition to attract investment prevents the emergence of continental-scale champions. Mario Pianta’s research on European industrial policy documents how the fragmentation of the European market has limited the scale of European firms relative to their American and Asian competitors, and his analysis explains why European champions in sectors such as automotive and telecommunications have struggled to match the scale of their rivals (Pianta, 2014). Anu Bradford’s work on the Brussels effect shows that the European Union can shape global regulation through the size of its market even when its firms are not the largest, and her analysis is a partial counterweight to the scale argument (Bradford, 2020). The European case demonstrates that a large population is a necessary condition for scale and is not a sufficient one, and the American case demonstrates that a market of continental scale is insufficient when the industrial base that would exploit it has been relocated.

The foreign policy consensus that governed American conduct since the Second World War is cracking, and the cracking is the development with the greatest consequences for the populations that the competition affects. The consensus rested on an agreement between the Republican and Democratic parties that domestic disagreements would not extend to foreign policy, and the agreement was reinforced by a funding structure in which the same corporate donors gave to both parties and ensured that neither was starved of resources (Wolff, 2026). Thomas Ferguson’s investment theory of party competition documents how campaign finance has shaped the convergence of the two parties on foreign policy, and his analysis explains why the consensus persisted for as long as it did (Ferguson, 1995). The agreement is now fracturing over Europe, where the Republican coalition has become willing to disengage from the continent and the Democratic coalition is moving in the same direction more slowly, and the fracture will become more pronounced over China. Andrew Bacevich’s critique of American military policy documents the pattern of intervention that the consensus sustained, and his analysis of how that pattern generated resistance at home and abroad anticipates the breakdown that is now visible (Bacevich, 2005). Stephen Walt’s analysis of the costs of the liberal hegemonic project makes a related argument about the gap between the ambitions of the consensus and the resources available to sustain them (Walt, 2018).

The trajectory that the current policy describes leads toward a confrontation whose consequences have not been adequately priced by any participant, and the inadequacy of the pricing is the most important observation that the analysis can offer. The United States and China each possess nuclear arsenals of a scale that would make a direct conflict catastrophic, and the Bulletin of the Atomic Scientists has maintained its Doomsday Clock at the closest position to midnight in its history, citing the combination of nuclear risk and climate breakdown that the current competition intensifies (Bulletin of the Atomic Scientists, 2025). The dominant ideology in the United States remains an assertive conviction that the country can shape global outcomes as it did in the second half of the twentieth century, and the conviction has not adjusted to the material balance that the analysis describes. John Mearsheimer’s structural analysis of great-power competition explains why a declining hegemon and a rising challenger tend toward conflict even when neither desires it, and his framework illuminates the mechanism through which the current trajectory reproduces the patterns of previous transitions (Mearsheimer, 2001).

The alternative that the analysis points toward is a negotiated settlement of the security architecture of Eurasia rather than of Europe alone, and the specification of the alternative is necessary even where the political conditions for it do not yet exist. A settlement of that kind would require the United States to accept that its capacity to dictate outcomes has diminished, and the acceptance would be resisted by the factions whose position depends on the continuation of the assertive posture. A settlement would also require the Chinese state to accept constraints on its own regional ambitions, and the willingness of the Chinese state to accept such constraints has not been demonstrated. Wolfgang Streeck’s analysis of the relationship between states and capital in the contemporary period explains why such settlements are difficult to conclude, because the disciplined state is one in which democratic pressures for redistribution and demilitarisation are structurally subordinated to the requirements of capital accumulation (Streeck, 2014). The conditions for a settlement are therefore not present, and the analysis can identify the direction in which a settlement would lie without predicting that it will be reached.

The distributional analysis of the programme identifies the populations that bear its costs and the populations that capture its gains, and the identification is the analytical contribution that a materialist framework makes to the discussion. The workers of the American industrial belt whose communities were hollowed out by the offshoring that preceded the Chinese rise are now told that the remedy for their condition is a confrontation with the country that replaced their employers, and the remedy does not address the condition. The workers of the Chinese export sector whose labour produces the goods that the competition is about work under conditions that the celebration of national champions does not describe, and their interests are represented by neither state. The populations of the Global South whose markets are contested by both powers have their development options constrained by the terms on which they are permitted to participate, and the people of the regions in which the proxy conflicts are fought bear costs that neither principal acknowledges. Jacob Hacker and Paul Pierson documented how the American political system has been organised to redistribute income upward, and their analysis explains why the costs of the current programme are being distributed in the same direction as the previous four decades (Hacker and Pierson, 2010).

The class analysis of the two systems must be conducted in terms of substantive outcomes rather than formal institutions, and the distinction is the one that a materialist framework requires. The American state is disciplined by corporate capital through campaign finance, lobbying and the revolving door, and Colin Crouch’s analysis of post-democracy documents how the formal institutions of representative government have become shells through which corporate interests are transacted (Crouch, 2004). The discipline has produced four decades of upward redistribution, decaying infrastructure and an industrial base that has been relocated. The Chinese state disciplines corporate capital on behalf of a national developmental project that it controls, and the discipline has produced the largest expansion of productive capacity and the largest reduction in extreme poverty in recorded history (World Bank, 2022). Neither configuration constitutes socialism in the sense of working-class self-emancipation, and the difference between them lies in which class the state is presently disciplined by rather than in the formal architecture of representation. The Chinese model’s limitations are real and are acknowledged by the Chinese state itself, which has identified the labour share, the household registration system and the social wage as problems requiring correction, and the direction of policy over the past decade has been toward addressing them.

The conclusion that follows from the analysis is not a prediction of which power will prevail but an assessment of the risks that the competition imposes on populations that have no voice in it. The financial structure of the American programme is fragile in the precise sense that its gains are concentrated and its losses would be distributed, and the constituencies positioned to profit from its collapse are not the constituencies that would bear its costs. The political constraints on the programme are tightening as the public absorbs its visible costs, and the constraints will bind more tightly as the costs become more apparent. The Chinese advantage in scale and scope is structural and will not be reversed by subsidies that address the cost of capital without addressing the structure of production. The foreign policy consensus is cracking in a way that could produce either a more rational settlement or a more dangerous confrontation, and the direction of the crack will be determined by the relative strength of the factions within the American ruling class and by the capacity of organised populations to impose constraints on both. The task that the analysis identifies is not the selection of a preferred power but the construction of the capacity of working populations in both countries to constrain the states that act in their name, and the construction of that capacity is the only development that would alter the trajectory that the evidence describes.

Authored By: Global GeoPolitics

Thank you for visiting. This is a reader-supported publication. If you believe journalism should serve the public, not the powerful, and you’re in a position to help, becoming a PAID SUBSCRIBER truly makes a difference. Alternatively you can support by way of a cup of coffee:

https://buymeacoffee.com/ggtv |

https://ko-fi.com/globalgeopolitics |

Bitcoin: 3NiK8BoRZnkwJSHZSekuXKFizGPopkE7

References

Amsden, A. H. (1989) Asia’s Next Giant: South Korea and Late Industrialization. Oxford: Oxford University Press.

Bacevich, A. J. (2005) The New American Militarism: How Americans Are Seduced by War. Oxford: Oxford University Press.

Bank for International Settlements (2025) Quarterly Review: Market Concentration and Financial Stability. Basel: BIS.

Bluestone, B. and Harrison, B. (1982) The Deindustrialization of America: Plant Closings, Community Abandonment, and the Dismantling of Basic Industry. New York: Basic Books.

Bradford, A. (2020) The Brussels Effect: How the European Union Rules the World. Oxford: Oxford University Press.

Bulletin of the Atomic Scientists (2025) Doomsday Clock Statement. Chicago: Bulletin of the Atomic Scientists.

Burry, M. (2025) Public comments on depreciation schedules for AI infrastructure, cited in Financial Times, 4 December.

Byman, D. and Pollack, K. M. (2007) Things Fall Apart: Containing the Spillover from an Iraqi Civil War. Washington, DC: Brookings Institution Press.

Congressional Research Service (2026) Semiconductor Fabrication Facilities Funded by the CHIPS Act: Project Status and Considerations for Congress. Washington, DC: CRS.

Crouch, C. (2004) Post-Democracy. Cambridge: Polity Press.

Ferguson, T. (1995) Golden Rule: The Investment Theory of Party Competition and the Logic of Money-Driven Political Systems. Chicago: University of Chicago Press.

Gilens, M. and Page, B. I. (2014) ‘Testing theories of American politics: Elites, interest groups, and average citizens’, Perspectives on Politics, 12(3), pp. 564–581.

Hacker, J. S. and Pierson, P. (2010) Winner-Take-All Politics: How Washington Made the Rich Richer and Turned Its Back on the Middle Class. New York: Simon and Schuster.

International Monetary Fund (2025) Global Financial Stability Report: Market Concentration and Systemic Risk. Washington, DC: IMF.

Mearsheimer, J. J. (2001) The Tragedy of Great Power Politics. New York: W. W. Norton.

Melman, S. (1974) The Permanent War Economy: American Capitalism in Decline. New York: Simon and Schuster.

Nolan, P. (2001) China and the Global Economy: National Champions, Industrial Policy and the Big Business Revolution. Basingstoke: Palgrave.

Pianta, M. (2014) ‘Technology and industrial policy in Europe’, in The Oxford Handbook of the Economics of the European Union. Oxford: Oxford University Press.

Streeck, W. (2014) Buying Time: The Delayed Crisis of Democratic Capitalism. London: Verso.

Teece, D. J. (1980) ‘Economies of scope and the scope of the enterprise’, Journal of Economic Behavior and Organization, 1(3), pp. 223–247.

Wade, R. (1990) Governing the Market: Economic Theory and the Role of Government in East Asian Industrialization. Princeton: Princeton University Press.

Walt, S. M. (2018) The Hell of Good Intentions: America’s Foreign Policy Elite and the Decline of U.S. Primacy. New York: Farrar, Straus and Giroux.

Wolff, R. D. (2026) US corporate dictatorship in crisis. Interview with Glenn, 14 November. Transcript.

World Bank (2022) Four Decades of Poverty Reduction in China: Drivers, Insights for the World, and the Way Ahead. Washington, DC: World Bank.



Leave a Reply