Global- Impact Newswire – For more than a decade, economists warned that the world’s next great inequality would not be income or trade, but the ability of nations to harness the engines of artificial intelligence. Now, a new World Bank assessment suggests that a huge divide has arrived faster, and more sharply, than anyone expected.
The bank’s Digital Progress and Trends Report 2025 released recently opens with a blunt admission: “Data on countries’ performance on AI chip ownership, data center capacity, and cloud adoption are scarce,” largely because the sector advances at breakneck speed, guarded by secrecy, and strained by geopolitics. Norms for measurement, it notes, are “lacking,” and the result is “fragmented and incomplete data availability on a global scale.”
But from what can be measured, one conclusion stands out. A few countries control nearly all of the world’s ability to compute. Everyone else is running behind.
The United States, the Bank finds, “has the highest raw compute capacity and the most-powerful supercomputers,” and it continues to dominate cloud markets, AI chip supply, and the supercomputing frontier. As of mid-2025, the country hosts “175 of the top 500 high-performance computing (HPC) systems,” accounting for about “50 percent of TOP500 supercomputing capacity,” and almost “40 percent of co-location data center capacity.”
Its tech giants such as Google, Meta, Microsoft, and XAI own “millions” of advanced AI chips, including NVIDIA’s coveted H100 units, “dwarfing the ownership of any other company in the world.”
It is the kind of supremacy that earlier eras expressed through steel, oil, or nuclear physics. Today, the currency is compute.
Europe’s Uneven Strength
Europe appears sturdier than most, yet its position is more fragile than it seems. The continent hosts “one-third of global HPC systems, capacity, and internet servers” along with 28 percent of data centers. Germany alone is home to “41 of the top 500 HPCs.”
But the Bank points out that the region lacks homegrown giants to compete with American cloud behemoths. OVHcloud, the largest European public cloud provider, recorded just $1 billion in revenue—“which pales in comparison” to Alibaba Cloud, Google Cloud, Azure, and AWS.
Europe also “lacks its own AI accelerator chips,” a reality that has sparked hand-wringing among policymakers eager to avoid dependence on American or Taiwanese semiconductor pipelines.
China’s Complicated Rise
China’s numbers tell a story of exceptional strength mixed with strategic vulnerability. It hosts “47 HPC systems, 9 percent of the world’s total,” and accounted for “13 percent of NVIDIA’s revenue” in 2024/25, though that share is falling because of intensifying trade restrictions.
Yet the country claims only “2 percent of supercomputing capacity,” a figure the Bank calls “likely an underestimate,” hinting that some Chinese systems may simply not be disclosed.
Even so, its global cloud exports, although small compared with the U.S., grew tenfold between 2016 and 2021on, one of the fastest expansions in the world.
The Vanishing Presence of the Global South
Beyond these hubs, the map goes dark.
Excluding China, low- and middle-income countries host “a mere 3 percent of the 500 HPC systems,” and just “1 percent of the supercomputing capacity,” despite housing 65 percent of the world’s population.
Nowhere is the void more visible than in Sub-Saharan Africa, which accounts for “16 percent of the global population” but “0 percent of HPC systems.” The region’s secure internet servers and data center capacity barely move the needle, just “1 percent” of the global share.
Low-income countries fare even worse: “0 percent of compute capacity.”
The Bank frames the imbalance starkly. High-income countries hold “86 percent of the top 500 HPC systems and a staggering 97 percent of the capacity.”
In practical terms, it means the regions likely to benefit the most from AI applications—health, agriculture, education, are the least able to build or train the systems that would power them.
“Governments have a pivotal role to play in building and strengthening these critical foundations so that AI can take root and flourish in developing countries,” says Axel van Trotsenburg, former Senior Managing Director at World Bank, in the report. “As the AI era accelerates, developing countries need to be prepared to leap forward. Small AI offers a unique opportunity to bypass traditional development barriers and spark homegrown innovation and inclusive growth.”
A New Kind of Dependence
Because compute is limited and costly, developing nations must choose between two paths: build local data centers, or rely heavily on foreign cloud providers.
But that dependency creates new vulnerabilities. The World Bank warns that international cloud reliance comes with “privacy and data sovereignty considerations,” and potentially exposes nations to “unauthorized foreign access, surveillance, and national security” risks.
Even the economics are complex. The United States contributes “87 percent of global cloud computing and data storage services exports,” the report says, and most of those exports go to wealthy countries. Low-income nations account for “0 percent.”
Among developing regions, Latin America shows the most momentum. Brazil and Mexico together account for “one-third of all US cloud computing exports to developing countries.”
But Africa barely registers.
Why Compute Investment Is So Hard
Building a data center is not like building a road or a port. It requires cheap electricity, reliable power grids, stable political environments, high-end technical skills, and the capital to buy AI chips that can cost “between $30,000 and $40,000 per unit.”
Even renting compute can be beyond reach. AWS charges “US$98 per hour” for a virtual machine powered by an NVIDIA H100, more than many African startups can afford for a full day, let alone months of model training.
These cost dynamics are intertwined with industrial concentration. The report notes that the compute supply chain is dominated by a tiny group of actors, each protected by network effects, scale advantages, and geopolitical leverage. As a result, “the market alone cannot ensure equitable access or distribution.”
A rare source of optimism appears in a single line: “Emerging models such as DeepSeek demonstrate that high AI performance can be achieved with lower compute demands.”
It is an understated but important point: efficient models may become the poor world’s most viable path to parity.
The Environmental and Security Toll
Compute is not just expensive; it is heavy. Data centers devour electricity and water, distort local housing markets, and burden infrastructure systems. Their environmental costs, the report warns, “are not fully reflected in service prices.”
Virginia’s booming data center corridor created “nearly 15,000 direct jobs” and more than “US$10 billion in economic output.”
But in Frankfurt, spiraling demand for “co-location data centers” helped push up commercial real estate prices, prompting the city to restrict where such facilities could be built.
For developing countries, where electricity costs are high and grids fragile,these trade-offs can be crippling.
Even when infrastructure is viable, the human capacity often is not. The report highlights “the scarcity of local technical expertise in cloud technologies and data center management,” compounded by chronic brain drain. Specialists who gain the skills necessary to run AI infrastructure often move abroad.
This, in turn, reduces investor confidence, creating a self-reinforcing cycle of underdevelopment.
The Path Forward
The World Bank stops short of prescribing a single strategy. Instead, it warns that compute must be treated as both an economic asset and a public good.
Governments must weigh whether “promoting local data center investment” is wiser than relying on regional hubs or global clouds. Any strategy, it argues, demands “careful planning, competitive neutrality, and clear sustainability strategies,” without which data centers risk becoming “costly, underused infrastructure.”
At its heart, the report issues a challenge to the global community. Compute is fast becoming the backbone of modern economies, but access is not evenly or equitably distributed.
And until that gap closes, the great AI revolution may deepen the divide between the world’s digital haves and have-nots.
Stay ahead in the world of AI, business, and technology by visiting Impact AI News for the latest news and insights that drive global change.
Got a story to share? Pitch it to us at info@impactnews-wire.com and reach the right audience worldwide!
Discover more from Impact AI News
Subscribe to get the latest posts sent to your email.
