The first major wave of cloud adoption asked a relatively simple question: which applications can we move out of the data centre?
The next phase is more complicated. Organisations are now deciding where each workload should run, which provider should operate it, what data can cross a border and how quickly the architecture can adapt when requirements change.
“Cloud 3.0” is useful shorthand for this shift. It is not a formal technical standard. It describes an operating model in which public cloud, private cloud, on-premise infrastructure, edge environments and sovereign services work as a coordinated portfolio.
The objective is no longer to move everything towards one destination. It is to place each workload where it makes the most sense.
AI is changing cloud requirements
AI workloads have very different infrastructure needs.
Large-scale model training may require centralised access to significant computing capacity. Inference, where a trained model responds to real users or systems, may need to happen much closer to the data, application or physical location.
For time-sensitive use cases, reducing latency can be critical. For regulated or sensitive applications, organisations may also need to control where information is processed and which legal jurisdiction applies to the provider, its administrators and its technology supply chain.
Research into emerging sovereign and specialist AI cloud environments points to growing demand for regional infrastructure capable of supporting AI factories, GPU services and local AI platforms. These workloads also place much greater pressure on network performance, resilience and the ability to scale across locations.
Sovereignty is moving into mainstream cloud strategy
Digital sovereignty was once viewed primarily as a public-sector or regulatory concern. It is now becoming part of wider infrastructure planning.
Gartner’s 2026 Market Guide for Cloud Infrastructure Sovereign Solutions says geopolitical tensions and advanced AI requirements are increasing global demand for greater control over proprietary data, infrastructure and core technology stacks.
The European Commission has also proposed a Cloud and AI Development Act that aims to triple data centre capacity in Europe over five to seven years. The proposal includes a common framework for assessing cloud and AI sovereignty, alongside measures intended to protect sensitive data and critical applications.
The proposal still needs to pass through the European legislative process, but the direction is clear. Cloud decisions are increasingly connected to resilience, industrial policy and control over critical technology.
Workload placement becomes a policy
In a Cloud 3.0 environment, placement should not be an informal architectural preference. It should be governed by clear policy.
That policy might take account of:
- Data sensitivity and residency requirements.
- Latency and performance.
- The cost and availability of specialist compute.
- Operational resilience and recovery.
- Portability between providers.
- The jurisdiction governing data, operations and administrative access.
The result may be a public cloud service for one workload, a private or sovereign environment for another and local edge infrastructure for real-time inference.
The difficult part is making those environments work together without creating fragmented security, duplicated tools or an architecture that becomes too complex to operate.
Technology providers need a more local view of demand
Cloud requirements will vary considerably by sector and geography.
A healthcare organisation may prioritise patient-data control and auditability. A manufacturer may focus on low-latency inference at a production site. A financial institution may want operational resilience across regions without losing control over privileged access.
That makes broad, one-size-fits-all messaging less effective. Technology companies need to understand which sovereignty, performance and resilience concerns genuinely affect their buyers, and where those buyers are willing to make trade-offs.
Cloud 3.0 is therefore as much a market segmentation challenge as an infrastructure one.
To understand how cloud, AI and sovereignty priorities vary across your target sectors or regions, contact Callum Budd at callum@mra-research.co.uk.
