A Race Defined by Hardware, Not Just Algorithms
When OpenAI CEO Sam Altman toured a vast construction site in Texas last month, the symbolism was hard to miss. Outfitted in work boots and a high-visibility vest, Altman was surveying a $60 billion data center project—an AI megastructure larger than Central Park, complete with its own natural gas plant. Once operational, it’s expected to rank among the most powerful computing hubs on the planet.
Thousands of miles away in Argentina, the disparity could not have been starker. At the National University of Córdoba, computer science professor Nicolás Wolovick oversees one of the country’s few AI computing hubs—housed in a retrofitted room strung with aging chips and tangled wires. His conclusion was blunt:
"Everything is becoming more split. We are losing." — Nicolás Wolovick, National University of Córdoba
This is the new digital divide—less about access to the internet and more about access to computing power. And it is redrawing the global tech map.
The Geopolitics of Compute
The chasm between nations that possess AI compute infrastructure and those that don’t is driving a seismic shift in global influence. According to research led by Oxford University, just 32 countries—roughly 16% of the world—host the large-scale data centers required to train frontier AI models. The lion’s share of these facilities are concentrated in the U.S., China, and the European Union.
The two dominant forces, the U.S. and China, now control more than 90% of all AI data center capacity used by businesses and institutions worldwide. These hubs, owned and operated by the likes of Amazon, Microsoft, Google, Tencent, Alibaba, and Huawei, represent not just technical capability but geopolitical leverage.
In contrast, more than 150 countries—including nearly all of Africa and much of South America—lack any meaningful AI infrastructure. India has at least five such centers; Japan, four. But even these are dwarfed by the scale of U.S. operations, which alone account for 87 AI computing hubs—about two-thirds of the global total.
AI’s Uneven Impact on Science, Language, and Security
The implications of this compute disparity go far beyond academic research. From healthcare to defense, AI is becoming a central driver of innovation—and those without access are being left behind.
Systems like ChatGPT perform best in English and Chinese, reflecting the linguistic environments in which they’re trained—namely, the countries with dominant compute power. Breakthroughs in pharmaceuticals, climate modeling, and genomics increasingly rely on massive AI workloads, inaccessible to countries without dedicated data infrastructure.
Even national security is affected. AI-enhanced weapons systems and surveillance tools are emerging faster in compute-rich nations, potentially shifting the balance of power on the battlefield.
Meanwhile, countries with no AI backbone are grappling with lost talent, stunted innovation ecosystems, and growing dependence on foreign tech providers. The digital sovereignty concerns are mounting.
"Oil-producing countries have had an oversized influence on international affairs; in an AI-powered near future, compute producers could have something similar." — Vili Lehdonvirta, Oxford University
Compute as the New Commodity
Much like oil in the 20th century, compute power is emerging as a strategic resource. Critical components like GPUs—particularly those produced by U.S. chipmaker Nvidia—are now key instruments in foreign policy.
In response, nations are adjusting. The United States and China have woven AI infrastructure into their trade and diplomatic strategies, with particular focus on regions like Southeast Asia and the Persian Gulf. Some countries are beginning to allocate public funds to develop domestic AI capabilities, seeking to avoid total dependence on external powers.
Smart Africa, a coalition working to harmonize digital policy across the continent, is among the organizations raising alarm.
"We have a computing divide at the heart of the AI revolution. It’s not merely a hardware problem. It’s the sovereignty of our digital future." — Lacina Koné, Smart Africa
The False Dawn of Digital Equality
The early 2010s brought optimism. Smartphones, cheap data, and expanding internet access helped level the digital playing field. App-based businesses flourished in developing economies—from mobile payments in Africa to ride-sharing in Southeast Asia.
By 2023, nearly 70% of the global population had internet access, up from just 33% a decade earlier, according to the UN’s International Telecommunication Union. It appeared that the world was converging digitally.
But AI has introduced a new layer of complexity. The infrastructure needed to run generative models and large-scale learning systems costs billions, putting it out of reach for most. The United Nations now warns of a widening gap unless significant interventions are made. According to their data, just 100 companies—mostly in the U.S. and China—accounted for 40% of global AI investment in 2023.
The digital divide hasn’t disappeared. It’s simply evolved—replacing bandwidth and mobile apps with teraflops and tensor cores.
Source: The New York Times, June 23, 2025.
