The product is compute. The business is the deployment system.

From research demonstrator to commercial infrastructure

Microsoft’s Project Natick demonstrated a full-scale, sealed subsea data-centre module that could be manufactured, deployed, operated on the seabed and recovered. Its engineering story is as important as the compute: a pressure vessel, seabed docking structure, shore power and data connection, marine installation method and planned retrieval all had to function as one system.

Commercial developments in China are now applying the modular principle at cluster scale. Public project material describes subsea cabins connected to shore facilities through power and data infrastructure, with marine deployment, remote operation and future expansion treated as part of the architecture. These projects do not remove the need for site-specific engineering; they make the complete delivery system easier to see.

• Sealed, retrievable modular compute systems
• Shore-side energy, switchgear, control and network interfaces
• Protected power, control and fibre routes
• Marine installation, ROV connection and commissioning
• Remote operations and a planned recovery pathway

Begin with the service—not the module

The starting point is the computing service the installation must sustain: required capacity, latency, availability, energy demand, security, operating life and recovery philosophy. Those requirements shape the module, but they also determine the shore landing, power architecture, fibre routes, redundancy and marine support model.

A delivery-led approach aligns the modular subsea data-centre system to a defined service need rather than treating it as a standalone product. It also defines the whole external operating system before major procurement commitments harden.

• Target computing capacity and availability
• Power demand, redundancy and emergency continuity
• Fibre connectivity, latency and route diversity
• Intervention, recovery and redeployment philosophy

Site and corridor readiness

The preferred location must be technically suitable and practically accessible. Water depth, seabed conditions, metocean environment, existing infrastructure, environmental constraints, shore proximity and vessel access all affect the delivery concept.

Site investigation should answer the questions needed by foundation or support design, cable routing, installation engineering and future recovery. The output is not simply a survey report; it is a decision-ready evidence set with limitations and open assumptions made visible.

Connect shore, power and fibre as one architecture

A subsea data centre depends on continuous connections. Power supply, conditioning, protection, telecom capacity, shore landing, cable manufacture, route engineering, installation and testing must be coordinated around the same design basis.

Interfaces that appear minor on an organisation chart can control the programme offshore: connector compatibility, bend limits, pull-in arrangements, test boundaries, spare strategy, shore works and the sequence in which systems become available. These decisions need one accountable interface register rather than separate supplier assumptions.

Design the marine delivery route

The installation concept must be buildable with available ports, vessels, lifting systems, handling equipment, survey support and weather limits. Module dimensions and weight influence much more than the offshore lift; they affect fabrication logistics, quayside readiness, transportation, seafastening, deck layout and contingency planning.

A credible programme takes the installation method far enough to validate asset availability, handling limits, cable sequence, testing, acceptance and fallback options before contracts are placed.

• Port and quayside capability
• Vessel, lifting and positioning requirements
• Cable installation and connection sequence
• Testing, commissioning and acceptance boundaries
• Weather, contingency and recovery windows

Operations begin during design

The operating model should not be left until installation is complete. Monitoring, remote diagnostics, cyber dependencies, condition thresholds, spares, access arrangements and decision rights must be designed alongside the physical asset.

Because the module is deliberately placed underwater, the intervention philosophy matters. Teams need to know which conditions can be managed remotely, when inspection is triggered, when recovery is justified and what assets or agreements make that recovery possible.

Prove one repeatable deployment system

The commercial opportunity is not a single successful installation. It is a deployment model that can be repeated without rebuilding every interface. That means capturing the design basis, supplier boundaries, vessel method, test evidence, operating data and lessons from the first live package.

Avenor’s role is to provide one route across technology supply and the external delivery system. Access opens the route to customers, sites and permissions. Assets bring together the modular compute system, vessels, cables, equipment and specialist capability. Assurance keeps design authority, interfaces, resilience, operations and recovery visible from the outset.