Hyperscalers are buying certainty, not land, and countries that understand this are winning the cloud-region race 

The contest to host global cloud regions is often described as a technology race. It is not. It is now a state capacity test.

Hyperscalers do not expand because ministers give conference speeches about innovation. They expand when a country can answer a far less glamorous set of questions with confidence: Can you deliver power at scale? Can you clear land and permits quickly? Can you guarantee regulatory stability for years? Can you provide dense fiber and low-latency interconnection? Can you show real demand, not just ambition? 

That is the real test. And it is becoming one of the defining economic questions of the AI era.

Across Asia and the Middle East, governments have begun to understand this with unusual clarity. India’s 2026 budget offered foreign firms using Indian data centers for global cloud services a tax holiday through 2047, explicitly to reduce uncertainty and attract investment. For a long time, a large share of India’s digital data, especially enterprise workloads, financial-platform backends, and early cloud deployments, was hosted primarily in Singapore, Europe, and occasionally the United States. Until the late 2010s, most global regions operated by Amazon Web Services, Microsoft Azure, and Google Cloud serving Indian users were located offshore, with Singapore acting as the dominant regional hub. 

Over the past decade, India has moved deliberately to bring more data and compute capacity onshore, using a combination of regulatory policy, public-sector demand creation, and infrastructure incentives. AWS, Microsoft and Google now operate more than a dozen cloud regions across India, compared with none a decade ago. The shift illustrates how coordinated policy signals can translate directly into hyperscale deployment at national scale.

Malaysia, meanwhile, has moved from aspiration to execution: Microsoft said in March 2025 that it would launch its first cloud region in Malaysia, with three data centers in greater Kuala Lumpur, nearly a year after announcing a $2.2 billion investment in the country. 

These are signals of how the game is now played. Countries that want cloud regions must stop thinking like digital marketers and start thinking like infrastructure operators.

During a recent Africa Hyperscalers Conversations session,  International Finance Corporation (IFC) Global Sector Lead for Data Centers and Cloud Investments, Obinna Isiadinso put the issue clearly. “Reliable electricity is the single most important constraint affecting data-center expansion in many emerging markets,” he said. “Without predictable access to energy, it becomes difficult to deploy infrastructure at hyperscale levels.”

Even mature markets such as Ireland and the Netherlands have introduced temporary data-center power restrictions as AI demand reshapes grid planning priorities. The implication is clear: countries competing for cloud regions are no longer judged only on connectivity or market size, but on their ability to deliver energy at scale and with certainty.

That observation matters because it strips away one of the most common myths in digital policy. For years, many emerging markets treated connectivity as the master variable. Build more fiber, land more subsea cables, expand broadband access, and the rest would follow. Fiber still matters, of course. But cloud-region decisions are increasingly shaped by what sits behind the fiber: grid readiness, transmission depth, redundancy, cooling, permitting, land assembly, and policy predictability.

Even the cloud providers’ own materials tell the story. Microsoft says Azure geographies are designed around data residency and compliance requirements, with business-critical apps kept on “fault-tolerant, high-capacity networking infrastructure.” Google’s region-selection tooling explicitly tells customers to weigh price, latency and carbon footprint. AWS says site selection takes account of environmental, geographic and regional regulatory risks. The subtext is obvious: hyperscalers do not choose locations casually, and they certainly do not choose them on rhetoric alone. 

That is why the countries now pulling ahead are not simply those with market size. They are the ones reducing execution friction.

India is a revealing case. The tax break is not important merely because it is generous. It is important because it addresses a specific investor fear: policy ambiguity. Reuters quoted Vaibhav Gupta of Dhruva Advisors saying, “This announcement helps in bringing clarity to foreign companies and lends stability in (their) tax position in India till 2047.” That is what hyperscalers buy before they buy land – certainty. India paired that tax clarity with a broader posture of cloud engagement, public digitization, and infrastructure signaling. Its IT minister, Ashwini Vaishnav, put the ambition more directly: “Data centres will be a major strength for India through which we can provide new services to the world.” 

The lesson is not that every country should copy India’s exact tax model. It is that investors reward coherence. When tax policy, public-cloud procurement, digital-sovereignty goals and energy planning move in the same direction, capital arrives faster.

Malaysia offers a different lesson. It has benefited from geography, especially proximity and low-latency links to Singapore, but geography alone does not explain the scale of inflows. Malaysia’s investment agency said approved investments hit a record RM426.7 billion in 2025, while Reuters reported that Microsoft’s Malaysia West cloud would begin operating in 2025. This is what cloud attraction looks like in practice: not one giant announcement, but a sequence of credible moves that convince global operators the market can absorb and support long-term deployment. 

This is where many African countries still misread the market. They assume the first hurdle is getting the cloud companies interested. In reality, the first hurdle is making the project financeable. That means pre-cleared sites, fast-tracked permits, power pathways, cross-border fiber resilience, and visible offtake. Hyperscalers may anchor demand, but they do not like walking into avoidable chaos.

Isiadinso made that point in the session as well. “Anchor tenants provide predictable revenue streams that reduce investment risk,” he said. “Where those conditions are present, projects become significantly easier to finance.” That is not just a financing observation. It is a policy instruction. Governments that digitize public services, commit credible cloud migration programs, and create sovereign hosting demand are not merely modernizing the state. They are underwriting bankability.

That distinction is crucial. Cloud regions tend to follow places where public policy creates a floor for demand and infrastructure planning creates a ceiling for scale.

So what must countries actually do?

First, they must treat hyperscale attraction as an inter-ministerial industrial project. The ministries for power, land, tax, telecommunications, environment, and trade must move together. Second, they must create long-duration policy visibility. Cloud infrastructure is too capital-intensive to be built on short political cycles. Third, they must prepare sites before courting investors, not after. Fourth, they must use the state intelligently as a demand shaper – through cloud-first procurement, digital identity systems, payment infrastructure, and local hosting mandates where appropriate. Fifth, they must build a workforce that can operate and secure these environments once they arrive.

This last point is often neglected. A cloud region is not just a real-estate decision. It is a talent decision. The countries that keep winning are not only supplying megawatts; they are supplying engineers, compliance professionals, security teams, and ecosystem partners. The best sites now come bundled with execution capacity.

Africa still has a real opportunity here, but only if it interprets the moment correctly. The continent’s opportunity lies in inference infrastructure –  the distributed compute layer that runs AI services close to users, financial platforms and governments. The continent should not imagine that the route into the compute economy begins with 1-gigawatt AI training campuses. In the IFC session, Isiadinso argued that Africa’s nearer-term opportunity lies in scaling enterprise, colocation, and regional compute layers before attempting to leap immediately to the most capital-intensive end of the stack. Countries move up the cloud hierarchy by proving they can host the next layer well enough to earn the layer after that.

The countries that will attract global cloud regions in this decade are not necessarily the richest, the loudest, or even the largest. They will be the ones that pass the hyperscaler test: power that clears, policy that holds, ready sites, networks that are dense, and demand that is real.

In the AI economy, cloud regions no longer go where the presentation deck is strongest. They go where the operating environment is hardest to argue against.