ThermAI Group Ltd
Explore FlightDeckFree GPU trial
Cloud GPU & CPU Compute

When The Compute You Have Isn’t The Compute You Need.

Cloud GPU & CPU compute for batch, parallel and schedulable workloads.

Bring a workload your current compute setup isn't serving well. We'll match its requirements to a supported trial and help you evaluate the results.

Free GPU trial Explore FlightDeck

Powerful Compute. Useful Heat.Our planned ThermAI Node network will reuse server heat to provide free heating and hot water for social housing residents. How it works

Conceptual blue fluid-flow visualisation
+ Start With Your Workload 01 / 03
Parallel Compute / Independent JobsIllustrative run
Independent tasks128 / batch
One workload. A clear comparison.
01 / Your Workload

Different Workloads.
The Right Compute Fit.

From recurring AI batches to independent scientific jobs and simulation studies. GPU compute is our focus, with CPU options where the application needs them.

Recurring GPU Workloads

AI & Batch Processing

Batch inference, model evaluation, fine-tuning and other repeatable GPU jobs. Test another compute option when recurring demand, GPU memory or capacity limits your current setup.

Compare your workload
Alongside Existing Resources

Research & Scientific Computing

Drug discovery, molecular dynamics, computational chemistry and independent scientific pipelines. Assess a portable workload alongside your existing cloud or cluster.

Assess a research workload
Exploring Fit Through Trials

Engineering & Simulation

Bring a CFD, simulation or design optimisation workload to assess. We’re testing fit case by case, starting with your solver, licensing and CPU or GPU requirements.

Discuss a simulation workload

Designed around queued and independent jobs. Continuous low-latency services and tightly coupled multi-node HPC fall outside our current proposition. Check workload fit

Independent and repeatable jobs Portable software environments Data that can be staged for the run
02 / ThermAI FlightDeck

Your Workload.
Meet FlightDeck.

Run GPU and CPU jobs, called Flights, with a Slurm-style CLI workflow. Explore the graphical version of that workflow in our interactive FlightDeck demo.

Explore FlightDeck

Run Flights With The CLI

Use the FlightDeck command-line interface for a Slurm-style workflow: prepare and submit jobs, then retrieve your results. We’ll help you set up access and run your first workload.

Supported pilot access

Explore The Visual Workflow

Prefer a GUI to a command line? Our interactive demo illustrates the graphical workflow, from selecting GPU or CPU resources to configuring a Flight.

GUI demonstration

Start With A Representative AI Workload.

Explore GPU and memory selection for a repeatable AI job. Bring your model, container and a representative batch so we can agree a useful comparison.

  • Batch inference
  • Model evaluation
  • Fine-tuning
Free GPU trial
ThermAI FlightDeckInteractive demo
FlightDeck demo showing GPU resource selection. Hardware shown is illustrative, not confirmed availability.
Explore GPU resource selectionIllustrative configuration

This GUI demo illustrates the FlightDeck workflow. Trial onboarding includes CLI access; we’ll confirm available resources and features with you. Demo configurations and estimates are illustrative.

01

Bring Your Environment

Discuss your container, dependencies and data.

02

Assess The Configuration

Check GPU, CPU and memory requirements.

03

Compare With Real Work

Agree what the trial needs to establish.

Trials Today

Supported onboarding using third-party cloud capacity. We agree the resources, hosting location and access before your first run.

Trial details
03 / Why ThermAI?

Keep What Works.
Test What’s Next.

You don’t need to move your whole stack to find out whether ThermAI has a useful place in it. Start with one representative workload alongside the cloud, cluster or hardware you already use.

01 / Identify

Start With A Specific Gap

Capacity, GPU memory or a workload that needs a different resource profile.

02 / Test

Compare A Real Job

Agree success measures for runtime, reliability, setup and total job cost.

03 / Evaluate

Make An Informed Next Move

Review the results together before agreeing any ongoing compute use.

One representative workload is enough to start the conversation.

Bring a workload to compare
Infrastructure / In Development

Powerful Compute.
Useful Heat.

We’re developing a decentralised network of ThermAI Nodes installed in homes. Each ThermAI Node combines liquid-cooled computing with ThermAI HeatBridge, designed to transfer recovered server heat into domestic heating and hot-water systems.

01 / Compute

Decentralised Compute

ThermAI Nodes will provide GPU and CPU capacity across a distributed network of homes. The network is designed for batch, parallel and schedulable workloads, with resources matched to each job.

02 / Recover & Reuse

Domestic Heating & Hot Water

ThermAI HeatBridge is designed to transfer recovered server heat into the home’s heating and hot-water systems, putting energy already used for computing to work a second time.

03 / Optimise

Smarter Energy Integration

Planned integration with solar, battery storage and smart tariffs is intended to optimise energy use, cost and carbon impact. We aim to pass operating savings on through lower compute costs.

Domestic heat recovery is part of the ThermAI Node rollout. Today’s cloud trials focus on validating compute workloads and workflows while we develop this infrastructure.

Social Value & Carbon

Less Wasted Heat.
More Value For Homes.

Our planned ThermAI Node rollout connects cloud compute with free heating and hot water for social housing residents. We’re building the evidence to help customers report the household and environmental benefits their compute supports.

01 / Social Value

Completely Free Heating & Hot Water

Our planned social-housing offer provides heating and hot water at no charge to residents, using waste heat from ThermAI Nodes. Compute demand helps fund a direct benefit for households.

02 / Carbon Measurement

Track Avoided Heating Emissions

We’re developing reporting that compares recovered heat with the energy and emissions of the heating it replaces, alongside the electricity used by compute.

03 / ESG Roadmap

Evidence Of Positive Impact

Our ambition is to offer verified carbon credits and social-value credits for customers’ ESG reports, supported by suitable methodologies and independent verification so customers can substantiate the impact they report.

Impact reporting and credits are in development. Avoided heating emissions will be reported separately from compute emissions; credits are not available today.

START WITH A REAL WORKLOAD

Test A Real Workload With A Free GPU Trial.

Bring work you expect to run again. We’ll assess the fit and agree what the trial needs to establish.

Up to £1,000in free compute credits

Eligibility, allocation, resources, timing and scope are agreed individually, subject to workload fit and availability.

Tell us about your workload

We’ll review it and get back to you about next steps.

Fields marked with an asterisk are required.

What are you looking for?

Tick the box below if you'd like to receive occasional updates from ThermAI. You can unsubscribe at any time.

To review your trial request, we need to store and process the information you provide.

We care about your privacy. Learn how we handle your data in our Privacy Policy.

How it works

  1. Tell us about your workload

    What runs today, and what you want to test.

  2. We’ll check the fit with you

    Review resources, software, data and delivery requirements.

  3. If it’s a fit, we’ll agree the trial

    Confirm scope, allocation and next steps.

Have a more complex requirement?
Talk to us at info@thermai.uk
A Few Practical Details

Compute Trials.
Clear Answers.

Check workload fit, understand the trial and know what to expect.

Ask us something else

Partners & Programmes

  • NVIDIA Inception programme
  • AIRSYS
  • OCF
  • 2CRSi
  • Oleon
  • University of Nottingham
  • Roedan
  • Inta
  • ESBE
  • Sunamp
  • InstaGen