AI, energy and sovereignty are reshaping the architecture of cloud computing

Executive Dialogue

Cloud computing was once presented as borderless and abstract. In the age of AI, that illusion is fading. Compute increasingly depends on energy grids, legal jurisdiction and physical infrastructure.

For more than a decade, cloud computing was framed as frictionless infrastructure. Compute appeared detached from geography, energy and industrial systems. Data moved globally while organisations increasingly treated cloud as an abstract utility layer.

Artificial intelligence is changing that logic. AI workloads require concentrated GPU clusters, stable electricity networks, cooling systems and secure data environments. What once appeared virtual is becoming physically visible again.

Cloud is no longer merely software infrastructure. It is increasingly becoming part of a broader industrial and geopolitical architecture.

The Physics of Privacy

The first era of cloud focused on scale and convenience. The emerging phase is increasingly shaped by governance, resilience and strategic control over data environments.

This shift is becoming particularly visible around enterprise AI, where the rush for innovation is colliding with the realities of data gravity, compliance and infrastructure dependency.

“The shift toward Private AI is a direct result of companies having moved data too quickly to the cloud.”

Charles Sansbury
CEO Cloudera

Private AI, hybrid architectures and controlled compute environments are no longer niche discussions. They reflect a broader reassessment of dependency, risk and long-term infrastructure design.

The question is no longer simply where workloads run — but under which legal, industrial and energy systems they operate.

The Thermodynamic Constraint

AI is re-materialising the cloud. Large-scale compute increasingly depends on:

• grid stability,
• power availability,
• cooling infrastructure,
• physical data centre concentration.

This changes how organisations think about scalability itself.

Cloud architecture is no longer governed solely by software efficiency. Increasingly, it is constrained by thermodynamics, energy policy and industrial coordination.

The cloud era is entering a more physical phase.

Architecture Returns

The traditional binary between “cloud” and “on-premise” is beginning to dissolve.

Hybrid infrastructures, edge deployments and private AI environments increasingly reflect a broader architectural transition: organisations are rediscovering infrastructure design as a strategic function.

“The question is no longer ‘cloud or on-premise’, but how to design a shared, efficient architecture.”

Sergio Gago
Global CTO Cloudera

The emerging model is not cloud-first. It is architecture-first.

The European Question

For Europe, this transition carries broader implications.

The majority of hyperscale cloud infrastructure remains under non-European corporate and legal control, while AI simultaneously increases the strategic importance of compute, data governance and energy coordination.

Europe increasingly defines the regulatory logic surrounding AI and digital governance, yet remains structurally dependent on external infrastructure layers to operate them.

If cloud is becoming infrastructure, sovereignty can no longer be treated as a secondary policy layer. It becomes architectural.

About Strategic Profiles
Strategic Profiles explore how infrastructure, AI and technological systems increasingly shape modern societies. This four-part profile examines how cloud computing is evolving from abstract software architecture into physical and strategic infrastructure.

---

Photo credit

Altair Media — Conceptual image generated with AI

Caption

Artificial intelligence is transforming cloud computing from abstract digital infrastructure into a physical architecture shaped by energy grids, data centres, legal jurisdiction and strategic control over compute capacity in Europe.