China and the United States: The AI and Robotics Revolution and Its Impact on the Rest of the World
Anthropic CEO Dario Amodei’s September 2026 statement that the pace of artificial intelligence development should be slowed to maximize its benefits and minimize its risks highlighted the rapidly developing technology’s impact on the world. In my own prior writing, I have examined its potentially revolutionary effects on global economic structures and dynamics, military affairs, political systems, and the balance of power among the United States, the People’s Republic of China (PRC), and other actors.
To date, much of the current discourse on the technological revolution unfolding before us, focuses on the two principal state drivers of the enabling technologies: the United Stated and the PRC, and the competition between them. This orientation was particularly highlighted was further showcased by the inclusion of Artificial Intelligence as a key theme in the September 2026 summit between U.S. President Donald Trump and his Chinese counterpart Xi Jinping.
Beyond Washington and Beijing, however, the potential impact of artificial intelligence on the rest of the world is less understood but equally significant. It will likely follow related, but distinct dynamics from those affecting the U.S. and the PRC.
At the end of September 2026, I had the opportunity to present my perspective on the subject at an event hosted by the Konrad Adenauer Foundation, the Institute for Strategic Dialogue and Research and Mexico’s Anahuac University. There, I engaged with fellow scholars in Mexico on the impact of artificial intelligence on “middle powers” and their role in the evolving international system. The present work presents my analysis on the subject, incorporating my remarks at the event, and insights from the associated interaction with colleagues.
Framing Concepts
Artificial intelligence is widely recognized as driving a multidimensional global transformation, with enormous, albeit uneven, and difficult to predict implications for global economic, military and strategic dynamics. Its full impact will come, however, through the incorporation of AI into robotics and other automation, in interaction with big data, ever expanding global connectivity, quantum computing, and other core technologies. These advances could produce enormous gains in productivity across a range of current and new industries, but they will almost inevitably also create existential risks. These include harms stemming from how the new technologies vastly expand the damage that humans can accidently or deliberately do to one another, including enabling genetically engineered novel pandemic-capable pathonagens and devastating cyberattacks, to triggering nuclear war, as well as the widely discussed “AI turns on their creators” scenarios.
Even as the adoption of and interaction between AI and other technologies produces enormous potential benefits, there is also a high probability that it will prove disruptive to global economic, military and security, and political dynamics. Along the way, it could increase inequality, criminality and unrest, weaken democracies relative to authoritarian systems, and shift the economic, military, and strategic balance of power toward China, as I have argued elsewhere.
The changes unleashed by these technologies may resemble the transformation produced by the Industrial Revolution and the rise of popular nationalism, yet they are likely to be faster, more disruptive, and potentially more deadly.
As part of this global transformation, middle powers ranging from Brazil and Mexico to Iran, Turkey, Germany, India, and others may face severe political and economic challenges, along with reduced autonomy and a shifting role in a transformed international system.
The pessimistic outlook for middle powers in this work reflects the likelihood that they will find it difficult to develop their own economically competitive frontier AI models and complimentary robotics and leading edge autonomous systems. They thus will be largely forced to choose between adopting the technology of dominant US and PRC-based firms, or have their economies and bases of national power become increasingly non-competitive with that of others who do.
As with prior technology and industrial revolutions, in the name of nationalism or autonomy, some political leaders will likely chose that latter self-destructive path of autonomy or anti-system behavior, just as some leftist governments in Latin America and elsewhere in the 20th century sought to pursue import substitution industrialization and other autonomous development strategies, ultimately at great cost to their relative economic position and national power.
The technology ecosystems between which middle powers will have to choose, will likely crystallize around those of the United States and the PRC. As with other elements of US-PRC strategic competition, in the interest of security and intellectual property protection, both governments will likely insist that with access to its most sensitive technologies, forgo integration with the ecosystem of their competitor.
Still, relevant if less capable alternatives may emerge around technologies developed by the European Union, Japan and South Korea, India, or some combination of these.
Despite such nuances and possibilities, the most likely course is that, in the coming technology revolution, middle powers and others are likely to have less space for autonomous development than they did during the era of the “non-aligned movement” during the Cold War. Indeed, such dilemmas may raise theoretical questions about what it means to be a “middle power” in the new environment.
⠀
Short Term Impacts on the Situation of Middle Powers
In the near term, AI-driven technological change will have significant economic effects on middle powers and other states, yet will likely have only a limited effect on their strategic position. For them, participation in the new technology economy will probably be dominated by initiatives by individuals and leading companies, rather than widespread and effective adoption of transformative technologies that could produce a systematic transformation of state capacity or the national economic base.
This is not to deny that significant cloud computing, artificial intelligence, and related projects are going forward among middle powers. In Latin America, Mexico and Brazil stand out as incipient hosts of cloud computing and some AI applications.
In Mexico, projects include more than $82 billion in current or planned cloud computing facilities, centered around an emerging cloud and AI cluster in Queretaro, although it is not yet clear that there will be adequate energy supply for all of the projects to go forward. South Korean, Taiwanese and Japanese companies have also committed to $7.7 billion in semiconductor manufacturing facilities in Guadalajara and Ciudad Juarez, mostly for chips destined for U.S. datacenters.
In Brazil, investors have committed to over $80 billion in data center projects, mostly in Ceará state, with about half of the promised funding coming from PRC-based ByteDance. Investors have also announced over $1.6 billion in semiconductor manufacturing facilities, mostly in Sao Paulo, Rio Grande do Sul and Amazonas states. Over $1 billion of that investment, the vast majority, is promised by the PRC-based firm Ziliatech.
Although the scale of such current investment sounds promising, from the perspective of the position of middle powers in the AI sector, the pattern to date seems to resemble neocolonial manufacturing relationships of prior eras, in which foreign capital financed the introduction of technologies and manufacturing facilities, whose operation and technology were largely controlled by those foreign interests, in partnership with local elites. Similarly, Mexico, Brazil, and other middle powers are courting Chinese, U.S., and other foreign investment of automation into factories, agricultural operations, or in the extraction strategic minerals, with little evidence of their own initiatives in this regard. This trajectory will make it likely that the foreign investors retain substantial control and capture much of the value added in these emerging arenas.
In addition to limitations in how middle powers are participating in the emerging new AI-data-automation economy, in terms of raw investment, AI and technology investment remains concentrated in the Great Powers. By comparison to the previously noted $80 billon investments in Mexican and Brazilian data centers, U.S.-based technology companies invested an estimated an estimated $460 billion in AI infrastructure in the country in 2025 alone, with an additional $600 billion anticipated in 2026.
For smaller countries, an emerging new option to harnessing natural resources to participate in the technology revolution is to “sell compute” like a natural resource. The Hive Digital Technologies project in Yguazú, Paraguay, and a proposed $500 million 10-megawatt sovereign data center, both leveraging leverage low-cost energy from the Itaipu dam for data storage and compute capabilities that can support external needs, are key current examples.
Even more than the previous generation of sales of commodity exports, such initiatives are likely to generate relatively little employment and may yield proportionally greater returns to foreign capital than traditional mining or oil ventures, as will be discussed later.
Beyond the question of the incorporation of AI, data, robotics and other automation at the state level, an important dynamic for middle powers is how the private sector and individual persons in middle and other powers are leveraging models and capabilities developed within the ecosystems of the great powers. Already, 40 percent of ChatGPT use occurs in middle-income countries. Adoption in lower-income countries remains constrained by limited access to computers, technical skills, reliable infrastructure and electricity, as well as corruption and instability.
Particularly in middle powers, advances in artificial intelligence are likely to impact different actors and institutions in the country unequally. A World Bank study found that private companies in middle powers have generally adopted artificial intelligence more quickly than governments.
For some states, artificial intelligence will afford the possibility to bolster the health of democratic institutions and public confidence in them. This includes tools which may enable greater monitoring of improprieties in government activities, including unexplained income and irregular activities of police, administrators, judges and other officials, thus helping to root out corruption. In an increasing range of cases, governments may use AI to recommend or make administrative or judicial decisions, rather than relying on fallible or corruptible human actors. Doing so may help restore the faith of people in the fairness of the system. Associated increases in administrate efficiency may boost perceptions of the utility of democratically elected governments.
On the other hand, use of such systems may also introduce errors which undermine faith in the system.
If artificial intelligence strengthen democratic governance in some cases, in others it may also facilitate authoritarianism. Governments already have access to vast amounts of data from cameras and other sensors, digital communications and financial records. AI increasingly provides those governments with the ability to identify patterns of concern from monitoring and cross-referencing that data in ways not possible with human analysts alone. It allows governments to identify not only potential criminal activities, corruption, or terrorism, but also activities, speech and suggested orientations that may be disfavored by the regime. AI may thus allow governments to follow and potentially act against individuals exhibiting undesirable although not illegal behaviors, dissenting opinions, or even disfavored sexual preferences and social attitudes. Through PRC-based companies and other pioneers of such technologies, middle powers and others will increasingly have access to such capabilities. Indeed, the recent controversy over “flock cameras” in the U.S. and their ability to monitor populations, however, illustrates that the ability of AI-empowered surveillance systems to empower state are not limited solely to Chinese technology or non-democratic governments.
Beyond impacts on democratic governance, the likely dominance of US and Chinese companies in new AI-enabled technologies, and decisions regarding their development will likely foment concerns and complaints in other parts of the world concerning a lack of control over them. Indeed, such concerns were recently surfaced by decisions by the U.S. to restrict access to Anthropic’s Claude Mythos 5 and Fable 5, impacting actors far beyond the United States.
Such concerns over AI will increasingly play a role in the competition between the US and the PRC and the relationship of both with other parts of the world. Beijing is already exploiting developing-country concerns about exclusion from AI decision-making to expand the position of its own AI models and AI governance. The July 2026 World Artificial Intelligence Cooperation Organization (WAICO) summit in Shanghai symbolized this effort, convening international technology executives and officials around cooperation and governance proposals that could indirectly favor the adoption of PRC open weight models and associated architectures and technologies in developing countries, at the expense of Western alternatives.
Transformation of the Geopolitical Environment
Although effects of the AI and automation revolution on middle powers through international and internal dynamics are already emerging, over the longer term, the position of middle powers will be shaped even more profoundly by changes in the geopolitical environment that the reinforcing effects of those technologies are likely to bring about.
Crime and Migrant Stresses. First, in the medium term, middle powers and other states may be affected by an empowerment of criminal and terrorist organizations brought by the new technologies.
Illegal armed groups in Mexico, Colombia, and elsewhere already use drones for surveillance, smuggling, and lethal attacks. Cryptocurrencies facilitate international money laundering and movement of funds, while new digital tools give small and nonstate actors greater capacity for theft, extortion, and cyber operations.
Middle and smaller states often lack the resources, technical expertise, institutional flexibility, and international cooperation required to counter drones, trace cryptocurrencies, or protect public and private infrastructure from cyberattacks.
Compounding the stresses on medium states due to crime and terrorism, Artificial intelligence will also strengthen the capabilities of larger powers such as the United States, the PRC, and the European Union to control their border and populations, making them more effective in blocking the flow of undocumented immigrants across their borders and locating and expelling those living in the country without legal status. Moreover, as discussed later, innovations in the use of robots and other automation in industrial countries for seasonal agriculture, restaurants, corporate cleaning and personal service work traditionally generating demand for immigrant labor in high-income countries, may compliment anti-immigrant sentiments already growing there, to economically sustain even greater restrictions on immigration than is currently the case.
While AI and other technology may theoretically expand the population management and border control capabilities of middle powers as well, the asymmetry between middle powers and major ones means that, in practice, it will likely be middle and smaller powers that will disproportionately be obligated to host displaced populations and receive deportees. These pressures will strain social systems and create additional opportunities for criminal organizations such as Tren de Aragua, Los Pulpos, and Los Gallegos.
⠀
Economic Displacement with Political Consequences. Beyond additional burdens from AI-empowered crime and migrant issues, over the medium term, artificial intelligence will likely generate enormous economic displacement with social and political consequences. Artificial intelligence and complimentary technologies such as robotics will create new jobs and increase productivity, but in the short and medium term it will likely eliminate positions more quickly than it creates them. Displacement will affect white-collar work in legal professions, health care, computer programming, systems administration, technical support, and customer service, as well as some forms of manual labor. The World Economic Forum estimates that AI alone could displace 92 million jobs by 2030. New jobs will not necessarily appear in the same places or require the same skills, creating significant friction as workers relocate and retrain.
Beyond job displacement, since artificial intelligence and complimentary technologies tend to be more capital than labor intensive, they will probably increase returns to those investing the capital, while putting downward pressure on wages in the categories of work they displace, particularly manual and lower-level technical jobs.
Unemployment and expanding wage inequality in middle income and other countries will likely be complimented by negative attitudes by some toward the political elites who are approving the deals that give technology intensive companies access to the country, from data centers and robotics-enabled manufacturing to large-scale projects in agriculture and mining. As with movements against foreign capital and local elites in prior eras, such perceptions, on top of real corruption and poor government performance, may fuel populist backlash against technologies perceived as the villains, along with the local elites aligned with them.
As noted in the previous sections, governments worried about maintaining order and attracting and protecting foreign investment in the context of such pressures, will be tempted to turn to AI-enabled social control tools, enabled by the same technologies, and perhaps even offered by some of the same foreign partners.
⠀
Accelerated Growth of PRC Strategic Advantage. Beyond pressures on their own governments, in the medium term, the position of middle powers may be impacted by the growing advantages of the PRC in the commercial position of its digital and automated technologies, as well as military advantages accruing to the PRC of commercial dominance in them.
In the realm of AI, PRC open-weight models such as the DeepSeek, Qwen and Kimi families, are designed to operate on less capable chips, cost less, and can be adapted more easily than many currently superior Western models. Those characteristics may help them win broad acceptance in developing markets.
The PRC is also deliberately integrating artificial intelligence with robotics and deploying those systems in factories, seeking the kind of global market dominance it achieved in electric vehicles and other industries. In robotics, U.S. products from companies such as Tesla and Boston Dynamics may have superior products and technologies in particular areas, yet PRC-based firms may ultimately prevail by producing at greater scale, allowing greater reductions in per unit cost, allowing them to gain even greater market share.
The proceeding analysis does not imply that the production of AI-enabled robots and other automated systems will not occur in middle-income and smaller countries. As with autos, PRC-based firms companies may manufacture units in partner countries, as it is already doing in Serbia, while retaining legal control of the design and other intellectual property, and realizing the lion’s share of return on capital on the investment.
PRC dominance of global commercial production of robots and other AI-enabled automatic systems would also give it a military advantage, within the PRC concept of civil-military fusion. A large PRC “automation industrial base” capable of producing tens of millions of advanced industrial robots, would give the PRC the capability to produce large numbers of robotic and unmanned systems for war, or even protecting its increasingly expansive global commercial operations. Indeed, the PRC is already exploring the utility of commercial robots from firms such as Unitree in public security roles.
As China’s power grows through these technology, commercial and military dynamics, it may act more assertively in regions where it has previously been deferential. This could include increasingly threatening countries that act against its corporations in Latin America and the Caribbean, such as it has recently done in Panama, or possibly being induced by its growing relative power over the west, to act decisively to end the autonomy of Taiwan.
⠀
The Dilemmas of Middle Powers in the New Security Dynamic.Middle powers will likely face growing dilemmas in this transformed geostrategic environment. Their relatively large populations will experience economic dislocation that expresses itself as demands on governments, or even unrest. Still, those governments will likely lack the capabilities of the great powers to fully exploit the potential of the new technologies on one hand, or to effectively manage the associated social and political pressures on the other.
In this new environment, the most consequential strategic choices of middle powers will likely concern alignment, and alternatively, disruption. A government may avoid formally declaring its political fealty to one power or another, yet its policy decisions to engage with the technological architecture of one great power, will probably oblige it to forgo the other’s technology ecosystem. The inherent choice will impact which of the great powers has more influence, with implications of economic partnerships and political alignment more broadly.
Given that US versus PRC technology environments will likely be built on very different frameworks regarding protection of the individual and their data, versus the empowerment of the state, choices by middle powers between the systems of the great powers will inherently affect the role of the state in their society and the economy.
As noted previously, some middle powers may seek to avoid such “choices” by pursuing a “non-aligned” path, attempting to build limited indigenous capabilities for autonomy. Faced with the limited options for economic competitiveness such a path affords, some states may chose a “disruptive” role, seeking international leverage through coercion, as Iran has done through its drone and missile program.
Long-Range Outlook for Middle Powers
Over the long term, as discussed in previous sections, middle powers will probably have less strategic influence and autonomy than they do today. As noted previously, they will generally lack full control over the AI models, robotics, and other autonomous systems they use. Their main leverage will come from negotiating entry into competing and increasingly incompatible ecosystems led by the PRC and the West, or from choosing a disruptive role.
One interesting question is the role middle powers could have in navigating or surviving major crises in the new order. Artificial intelligence and the instability of global transition could produce disasters with worldwide, even existential, consequences. These might include AI-engineered pandemics, cyberattacks that disable the global economy and life-sustaining infrastructures, or the accidental or deliberate escalation of nuclear conflict. Amidst such crises, middle powers may possess sufficient resources, scale, and independence to mediate between great powers, thus altering the course of events. Alternatively, they may be best positioned to emerge as survivors following such an Armageddon.
Short of such cataclysms, the transformed nature of the new technical and economic order may create unique new strategic roles for small and medium states.
Over the long term, a key question for the incorporation of middle and smaller powers into the new system will be their bargaining power, in the previously discussed emerging dynamic as sites for extracting “compute” as a resource, or in hosting factories, mines, or agricultural projects enabled by AI, robotics and other automated systems.
Unlike earlier industrial development, which requires good transport infrastructure and a sizable skilled workforce, the generation of “compute” depends primarily on abundant electricity, the connection of electricity and data to the compute facility, and the importation of significant quantities of high-value equipment. A data facility can be built by the state or directly by foreign investors and operated by relatively few technicians. Although AI may produce substantial compute revenue to the firm operating it, and royalty payments to the government permitting it, such technology enclaves may not generate significant local employment, supply-chain benefits, or broader economic development.
Within the dynamics of selling compute as a type of new resource colonialism, the competition for data-center and other investment related to the new technologies could produce a race to the bottom among small and middle powers. As with today, governments may offer inexpensive electricity, political guarantees, favorable taxes, and permissive legal treatment to attract multibillion-dollar projects.
In negotiating technology intensive investments like compute or factories running on AI-enabled robotics and automation, the balance of power will likely strongly favor the foreign investors over the countries. By contrast to investment in mines and factories, which are difficult to remove once built, and possible to realize its value after expropriating, much of a data center’s value lies in equipment that can be relocated, becomes obsolete quickly, and is difficult for an expropriating government to operate. Sensitive equipment could also be disabled or destroyed by a technology company to protect intellectual property before a government or others could effectively expropriate and exploit it.
Even more than today, the negotiation for massive foreign technology investment may generate serious risk of “elite capture.” Access to the country by technology firms is an extremely valuable permission, based on decisions by a small number of local decisionmakers, yet having only limited impact on the broader economy. In countries with weak institutions and a tradition of corruption, the incentive will be great for foreign firms to secure and protect access through bribery and other particularistic benefits for elites. These countries will be especially vulnerable to the technology firms of foreign powers such as the PRC which have a weak record of government enforcement of bad behavior by its firms abroad. Those foreign firms and the elites they capture, in turn, may find it more cost effective to invest in maintaining control, rather than investing in the broader community, in order to sustain social support for their project.
Although the sale of “compute” may present lucrative options for middle and lesser powers, they may also use their physical resources and geography to generate value from the new technology in more traditional ways ways.
As mentioned previously, over the long term, AI-enabled automation may have its greatest effect in middle powers through not only data centers and robot-filled factories, but also agriculture and extractive industries. Here, advanced machines will perform planting, spraying, irrigation, maintenance, harvesting, logistics, mining, and transport. These applications could increase output while displacing workers from sectors that have traditionally sustained marginalized populations. It could also increase economic pressures on middle powers by decreasing demand in developed nations for seasonal agricultural workers.
In order to attract technologies, or due to elite capture, some states could adapt their political and legal systems to become enclaves for technology companies or even be transformed in their entirety into technology experiments. Current discussion in Argentina about treating AI-run companies as legal persons, foreshadows what could be possible.
Beyond the changing role of states in the global system through new technologies, middle and smaller powers, another important group of questions centers on how middle powers and smaller states incorporate technologies within their own societies.
In middle and lower income countries in particular, wealthy households may use robots for personal service and security, amplifying existing class divisions. Still, relatively low wages in those countries will continue to make human domestic labor viable there, in many settings.
In some middle and smaller countries, social preferences may compliment such effects, slowing the adoption of robotics for personal service. As currently happens with the preference of human domestic servants over mechanical appliances such as dishwashers in many more traditional societies, the role of humans for some tasks may be preserved, even where machines are less expensive, owing to a preference in the population for human interaction.
Conclusions
The transformation of global economic, military, and political systems through advances in artificial intelligence, robotics, and their interaction analyzed by this work is only beginning. The dynamics hypothesized here are intended to stimulate thinking about what may occur and how governments can manage the resulting challenges.
Whatever the veracity of the specific developments anticipated in this work, its central argument is that emerging technologies will affect middle and lesser powers differently in in some areas, than they will affect the great powers. Those differential dynamics, in turn, will impact how middle powers and smaller states relate to the US and the PRC, and how the latter interact with each other.
The future is never predetermined. Nonetheless, it is vital for political leaders in middle and lesser powers must think carefully about how emerging technologies will affect their countries, positioning their societies to seize opportunities and manage risks, ensuring that they are not merely passive recipients of a transformed world, but instead, constructive participants in shaping it for the benefit of their own people and the broader international community.
