One of the most consequential economic and geopolitical contests of the modern era is taking place between American and Chinese technology companies competing to develop artificial general intelligence (AGI), a technology that could fundamentally alter the relationship between capital, labour, and economic growth, and in doing so change the balance of power between nations.

The ambition is to develop computer systems capable of performing most of the intellectual tasks that humans can perform, from analysing financial markets and writing software to conducting scientific research, providing legal advice, and managing complex business operations. Existing artificial intelligence systems can already perform many of these functions to varying degrees. General AI would extend those capabilities across a much wider range of activities, allowing machines to reason through unfamiliar problems, acquire new knowledge, and apply that knowledge independently. No universally accepted definition of AGI exists – it is one of those things that is hard to define but will be easy to recognise – nor is there agreement about how close its development might be, but the economic consequences of achieving it could be extraordinary.

The leading competitors include America’s OpenAI, Google DeepMind, Anthropic, Meta, and xAI, alongside Chinese developers such as DeepSeek and Alibaba. Their competition extends beyond the sophistication of their software to the vast computing infrastructure needed to develop and operate it. Advanced semiconductors, data centres, electricity generation, engineering expertise, and access to capital have become strategic resources in a contest whose outcome could determine the technological leadership of the coming decades.

American companies enjoy substantial advantages in capital investment and access to advanced computing infrastructure, while Chinese developers have demonstrated an ability to build increasingly competitive systems using comparatively constrained resources. The possibility that Chinese companies could overcome American technological advantages through improvements in efficiency adds another dimension to the contest.

Historical Context

In a brilliant essay published in Foreign Affairs this week, entitled Empires of AI: The Rise and Fall of a New Order, historian Niall Ferguson, researcher Lisa Klaassen, and former Google chief executive Eric Schmidt place this competition in a much wider historical context. Their central argument is that artificial intelligence should be understood as an imperial technology, with the United States and China emerging as rival technological empires.

The comparison is important because great powers have historically established and sustained their international dominance through advantages in technology, productive capacity, infrastructure, and access to capital. The development of AGI could create a new foundation for such dominance, allowing the countries that control advanced intelligence systems to exercise considerable influence over the economic and technological development of other nations. The argument also raises the question of how durable such dominance might prove. History suggests that technological advantages seldom remain the exclusive preserve of one power indefinitely, and that the spread of those advantages can contribute to the rise of challengers and the decline of established powers.

The economic significance of AGI lies principally in its potential to reduce the cost of intellectual labour. The industrial revolution mechanised physical production, while the computer revolution transformed the processing and distribution of information. General AI could go considerably further by automating much of the reasoning, analysis, planning, and problem-solving on which modern service economies depend. A company employing hundreds of accountants, analysts, software engineers, administrators, and legal specialists might eventually perform a substantial proportion of their work with a much smaller workforce supported by advanced AI systems. The resulting productivity gains could be large, allowing businesses to produce more output at substantially lower cost.

New industries and economic activities could emerge as services that were previously prohibitively expensive became widely affordable. But the accompanying disruption to employment could be equally significant, particularly for workers whose economic value depends on intellectual skills that have historically commanded high wages.

Geopolitical Importance

These possibilities explain the geopolitical importance of the competition identified by Ferguson and his co-authors. A country whose businesses successfully deploy general AI could acquire considerable productivity advantages over its competitors, while accelerating scientific discoveries in medicine, engineering, energy, and advanced manufacturing. Military planning, intelligence analysis, cybersecurity, and weapons development could similarly be transformed. Countries that depend on foreign AI systems could consequently find themselves dependent on the technological infrastructure and commercial decisions of the powers controlling those systems.

The contest between Washington and Beijing is therefore about considerably more than the commercial revenues of the technology industry. It concerns the possibility that control over advanced intelligence could become one of the principal foundations of international economic and strategic power.

For financial markets, however, the implications are more complicated. Enormous amounts of capital are being committed to semiconductor manufacturing, data centres, energy infrastructure, and the development of increasingly sophisticated AI models. Those investments are being made in anticipation of substantial future economic returns that have yet to be demonstrated on anything approaching the scale implied by the ambition.

If AGI produces a major acceleration in productivity, the resulting gains in corporate profitability and economic growth could justify much of the investment. If progress proves slower, the technology more expensive to operate, or the commercial returns more widely distributed than investors anticipate, considerable amounts of capital could have been misallocated. It is entirely possible that general AI eventually transforms the global economy while many of the companies financing its development fail to capture a proportionate share of the resulting wealth.

Potential Consequences

For South Africa, the potential consequences run in both directions. Affordable access to advanced intellectual capabilities could substantially improve the competitiveness of businesses constrained by shortages of skilled workers, while opening opportunities to improve education, healthcare, engineering, and public administration. Smaller firms could acquire capabilities previously available only to large organisations. Yet widespread automation could also weaken demand for sections of the skilled workforce, creating additional employment pressures in an economy already struggling to generate sufficient jobs.

The geopolitical dimension is equally important. Given the general aversion to merit, competitiveness, and investment that runs through the bulk of South African government policy (just consider its rejection of Starlink as a most basic example) South Africa is unlikely to develop frontier AGI systems independently and will therefore depend heavily on technologies developed abroad. The decisive question will be whether the country can use those technologies to achieve substantial productivity gains while avoiding excessive dependence on any single foreign technological power.

Beyond these immediate considerations lies a possibility of still greater consequence. If advanced AI systems become capable of conducting independent research and improving the design of subsequent AI systems, technological development itself could accelerate.

Scientific and engineering problems that currently require years of work might be resolved considerably faster, potentially compressing decades of innovation into much shorter periods. Whether that becomes possible remains uncertain, as do the practical limitations and risks associated with increasingly autonomous systems. Nevertheless, the implications are sufficiently important to explain the extraordinary intensity of the present competition. The battle for general AI is ultimately a battle over the future cost of intelligence itself, and therefore over the distribution of economic wealth and power in the decades ahead. The countries that master its development and application could acquire advantages comparable to those enjoyed by the great industrial powers of earlier centuries, while those that fail to adapt risk becoming increasingly dependent on the technological capabilities of others.