AI factories.
Any climate.
Half the cost.

Contact us

The right way to build your data center.

Mission

We build scalable and automated AI infrastructure.

A more sustainable AI future. Anywhere.

The problem

Conventional infrastructure was built for yesterday’s compute.

AI workloads demand more power, more density and faster deployment, but traditional data centers are slow, expensive and resource-intensive.

Years to deploy

Traditional data centers take years to plan, permit and build.

Capital intensive

High upfront infrastructure cost and increasing operational expense.

Water and location constraints

Conventional cooling can limit where dense AI capacity can be built.

Our solution

A building-free platform for next-gen AI infrastructure.

Integrated power, immersion compute, dry heat rejection, service stations and automation in a repeatable architecture.

DC Uptime platform with site power, two service stations, two immersion compute trenches, two dry cooler arrays and rail-based robotic maintenance Close-up of a DC Uptime immersion compute module with submerged servers and internal heat exchanger
01

Site Power & Network

Utility power and fiber connect directly to the platform.

02

Service Station

Power conversion, controls, network aggregation and maintenance hub.

03

Dry Cooler Array

Closed secondary loop with water-free heat rejection.

04

Automated Maintenance

Rail-based robotic service integrated with each compute trench.

05

Power & Network Distribution

Power, network and controls distributed through redundant pathways.

06

Immersion Compute Modules

High-density compute in independent trench-integrated immersion modules.

Automated maintenance

Rail-based robotic maintenance.

A rail-mounted system extracts, transports and services immersed servers so routine maintenance is built into the platform.

  • 01Automated extraction and insertion of wet servers/modules
  • 02Transport to the service station
  • 03Reduced routine field handling
  • 04Repair and diagnostics offline at the service station
Rail-based robotic system extracting a server module from an immersion trench
Automation keeps AI online.

How it works

From site to AI in six steps.

A repeatable module-level design from civil work to automated service.

1
Excavated trench

Excavate trench

2
Immersion IT module inserted into trench

Insert immersion IT module

3
Compute immersed in dielectric fluid

Compute immersed in dielectric fluid

4
Heat exchanger transfers heat to water-glycol loop

Transfer heat to water loop

5
Dry coolers rejecting heat to ambient air

Reject heat with dry coolers

6
Robotic service system

Service with robotics

Economics built for scale.

500 kW+ IT load per module
PUE as low as 1.02 High-efficiency thermal architecture
Zero water No water consumption for cooling
Lower cost To build and operate
$3M / MW Turn-key deployment of compute modules with integrated cooling, power distribution, and automated maintenance.
Scalable
efficient
resilient
for what's next

Preliminary estimates. Subject to change due to customer scope, site conditions, redundancy and commercial terms.

Traction

Real progress.
Strategic partners.

Strategic partner

ITS Infocom

Signed MOU and pilot deployment discussions with a data center developer, owner and operator.

Commercial validation

Pilot discussions underway

Engagement with data center owners and operators around initial deployments and commercial validation.

Get in touch

Let’s build the future together.

Interested in learning more, partnering or investing? We’d love to hear from you.

Thank you.

Your message has been sent. We’ll get back to you shortly.