Global hyperscalers gave governments, banks, startups and service providers access to modern infrastructure faster than anyone could have built it. That deserves saying plainly before anything else.
Then came the question that follows every piece of infrastructure a country relies on: should a nation’s digital future depend entirely on platforms controlled somewhere else? Increasingly, the answer is no. The standard response has been a local region of a global platform, described as sovereign. The data stays in the country. The control plane, the operations and the roadmap do not.
So countries are turning to national cloud providers: telecom operators, data centre operators and national technology companies running cloud on home soil. This article is about their side of it, because reducing dependency turns out to be less about building a cloud than about running one people choose.
Three kinds of dependency
Dependency on a global platform takes three forms, and a national provider has to answer all of them.
- Control. Who operates the systems that government, health, finance and emergency services now run on, and under whose authority.
- Jurisdiction. Whether citizen data and sensitive workloads can be governed by national policy, and whether that can be verified rather than assumed.
- Value. Where the spending goes, and whether local providers, integrators and engineers get a share of it.
A region solves part of the second. A national provider has the chance to answer all three.
Not a replacement. A real alternative.
National cloud is often misread as a campaign against public cloud. It is not, and the providers that succeed do not pitch it that way.
Most organisations will keep running hybrid and multi-cloud estates. Global applications and rapid experiments belong where elasticity earns its premium. Regulated, sensitive and steady workloads belong where control, residency and predictable cost matter more. The national provider’s job is to be the credible home for the second group, so that placement becomes a decision rather than a default.
That is also what resilience looks like. Dependence on any single provider is a risk, whoever the provider is. A country with a real domestic option has more choice, more redundancy and a stronger hand in every negotiation.
Chosen, not mandated
Here is the part national cloud programmes most often underestimate. A mandate can move the first workloads. It cannot keep them. A national provider reduces dependency only when ministries, banks and enterprises would pick it anyway.
That means offering what customers now expect from any cloud: self-service, compute, storage and networking, Kubernetes, managed databases, backup and disaster recovery, security services and GPU capacity. It means pricing a finance team can forecast, support from engineers in the same time zone, and compliance a customer can evidence to its own regulator. A national cloud has to operate as a business, not as an infrastructure project with a flag on it.
A mandate can move the first workloads. Only the service keeps them.
One country, many boundaries
A national provider does not serve one kind of customer. A ministry handling classified material, a hospital group, a central bank, a university and a startup all need different boundaries around their data.
That is why the deployment model matters as much as the data centre. A shared national cloud lets many public tenants run side by side on one substrate, each with its own keys and identity. A private government cloud serves a single body that needs infrastructure of its own. Air-gapped environments serve classified work. A provider that can offer all three from one platform can serve the whole country without running three businesses.
Do not swap one dependency for another
There is a quieter risk. A national provider can leave the hyperscaler and land in a different dependency: one infrastructure vendor’s licence, one proprietary stack, one supplier whose roadmap now shapes the national cloud.
This is why so many national initiatives build on open and interoperable technology, and why the platform layer should be infrastructure-agnostic. A national cloud that can run above OpenStack, CloudStack, Kubernetes, VMware or bare metal, on hardware from the vendor it already buys from, keeps the freedom to change course.
The same applies to operations. Some providers will run the platform entirely with their own teams. Others will have a partner operate it for them. Either can be sovereign, as long as the operator answers to national law and the provider decides.
The catalogue gap
Every national provider meets the same objection. Customers compare it with platforms that offer thousands of services. A cloud built for one country can only offer what one country can build and run.
On its own, that is true. It is also why so many national programmes end up accepting the trade they set out to avoid: a cloud that works, or a cloud that is theirs.
The answer is not to become a hyperscaler. It is federation. With Oasis Stack, a national cloud keeps its customers, data, operations, identity and keys inside its boundary, and by explicit agreement offers and consumes services with other independent clouds. Customer data never moves automatically. The catalogue grows beyond what one provider could build alone, and the control stays home. Sovereignty should not cost you scale.
The platform can be run by the provider’s own teams or operated by CloudHost, with support in three layers, and it launches in 90 days rather than the years a platform built from scratch takes.
What successful national providers share
- An anchor, not a crutch. Government support gives direction and early adoption; the service has to earn the rest.
- An ecosystem, not a monopoly. MSPs, software vendors, integrators and security firms selling through the national catalogue.
- Governance customers can prove. Controls that are enforced, and evidence that is produced as a matter of course.
- A business model that lasts. Revenue from infrastructure, platform services, security, AI and the marketplace.
The goal was never to remove global cloud from a country. It was to make sure the country has a choice. The national providers that deliver one will not simply host a digital economy. They will own a real part of it.