Distributed sovereign compute infrastructure

Compute, right-sized.

MicroScalers are modular, locally anchored compute nodes that bring AI and digital infrastructure closer to the places that need it — integrating compute, energy, storage, cooling, water, heat and connectivity as one resilient system.

Reference designs typically target regional-scale nodes, including 1–5 MW configurations, with capacity adapted to site and workload.
Energyrenewables + storage
Water & coolingclosed-loop / low-water
Local valueheat + grid services
Connectivityfederated + resilient
MicroScalersovereign compute node
The concept

A new infrastructure layer between enterprise edge and hyperscale cloud.

A MicroScaler is a compact, modular compute facility designed as part of the local infrastructure system around it. Multiple nodes can operate as a federated network, creating regional compute capacity that can grow in smaller steps and align with local power, water, heat, land and connectivity conditions.

Working definition
MicroScaler = an individual distributed compute node.
MicroScalers = the network and infrastructure category.
01 — DISTRIBUTED

Regional by design

Capacity can be placed closer to demand, critical infrastructure, industry, research clusters and communities instead of concentrating every workload in a small number of very large sites.

02 — SOVEREIGN

Control with portability

Data location, workload placement, operational control and technology exit paths become explicit design choices. Portability is built into the architecture rather than treated as a future migration problem.

03 — PRODUCTIVE

Designed for local value

Power, storage, cooling, recovered heat, water and grid services are planned together, enabling the compute node to participate in the wider energy and infrastructure system.

Integrated infrastructure

Water. Power. Data. And the flows around them.

Digital capacity is becoming a third essential infrastructure layer alongside water and power. MicroScalers treats the physical and digital flows as a single design problem, so site selection and operations can optimize resilience, resource use and local benefit together.

◈
ComputeAI, HPC, cloud and regional workloads
ϟ
Powerrenewables, grid supply and microgrids
▣
Storagebatteries and flexibility services
◌
Coolingliquid, closed-loop and low-water systems
≈
Watersite water balance and resilience
↗
Heatproductive reuse where demand exists
Resilience

A network can fail differently from a single giant site.

Federated regional nodes can support continuity, geographic redundancy and staged capacity growth. The design goal is resilient service across the network, supported by local energy, storage and multiple connectivity paths.

ContinuityWorkloads can be distributed across nodes and locations.
ModularityCapacity can be added in smaller increments as demand grows.
Local energyStorage and renewable generation can support site and grid resilience.
Sovereignty

Operational sovereignty is more than data residency.

The architecture can combine European data and model governance with practical control over compute, orchestration, location and migration. Current accelerator ecosystems can be supported while preserving planned alternative back-end routes and exit options.

DataControl over where sensitive information is stored and processed.
ComputeRegional capacity for public, industrial and strategic workloads.
PortabilityWorkloads and models designed for movement across infrastructure layers.
Where MicroScalers fit

Compute infrastructure designed around real places.

The model is relevant where digital demand intersects with energy constraints, resilience needs, sensitive workloads, available heat sinks, industrial infrastructure or local development objectives.

AI

Regional AI & sovereign cloud

Local inference, model serving, research computing and sensitive workloads with clear jurisdictional and operational control.

⚙

Industry, ports & brownfields

Compute integrated with industrial energy systems, storage, circularity, heat demand, logistics and redevelopment.

+

Critical & civic infrastructure

Hospitals, utilities, public services and regional systems that benefit from continuity, proximity and resilient local capacity.

⌂

Campuses & innovation districts

Universities, research parks and mixed-use districts combining compute demand with energy flexibility and heat reuse.

△

Agri-energy clusters

Smaller nodes can be explored alongside farms and controlled-environment agriculture where heat, power, land, water and income streams align.

◎

Resilient regional networks

Multiple standardized sites can form a federated infrastructure layer across cities, regions and countries.

Site intelligence

From data centre planning to infrastructure digital twins.

Site selection can combine satellite and building data with power, water, connectivity, climate exposure, heat demand, land constraints and ecological indicators. The same data layer can support ongoing measurement of performance and resilience.

SITE

Understand the place

Assess buildings, roofs, land, energy, connectivity, water conditions, physical climate risk and potential heat users.

satelliteweathergridwater
MODEL

Design the system

Match compute demand with generation, storage, cooling, water constraints, heat recovery, resilience and local infrastructure.

computeBESScoolingheat
MEASURE

Track real outcomes

Connect operational data with energy, carbon, water, resilience, property and ecological indicators to support transparent reporting.

MRVresiliencelocal value
The aim is productive digital infrastructure: compute capacity that strengthens the energy, water and economic systems around it while creating durable regional capability.
Development principles

System quality defines a MicroScaler.

Resource-aware

Design around local grid conditions, renewable supply, water stress, cooling requirements and climate resilience.

Community-aligned

Create tangible local value through heat, jobs, infrastructure services, education, research and regionally relevant applications.

Technology-portable

Use modular hardware and software layers that preserve operational choice and reduce long-term dependency on any single stack.

Measurable

Track energy, water, carbon, resilience and local outcomes with auditable data and clear performance indicators.

Replicable

Standardize the parts that should repeat while adapting each site to its physical, regulatory and economic context.

Financeable

Combine contracted compute demand with infrastructure revenues such as energy flexibility, heat and resilience where viable.

In development

Building the reference architecture and first demonstrators.

MicroScalers is an emerging infrastructure concept. Current work focuses on reference architectures, research collaboration, site assessment, demonstrators and financing models across European deployment contexts. Research teams, case studies and project updates will be added as the initiative develops.