Physics World Models for the Engineered World

KYAMOS

Holy Grail of Multiphysics Simulations

TUNet-AI is our proprietary, time-aware artificial intelligence for fast multiphysics prediction. It brings together time, space, geometry, materials, boundaries and scale through dimensionless numbers for computer-aided engineering and demanding engineering applications.

80
GPU dies across 10 nodes
11
Employees
€2.3M
Competitive R&D funding
14
Funded HPC & AI projects
Time × space × multiphysics

TUNet-AI: physics-aware prediction

Physical context meets fast field prediction. These are the capabilities at the heart of our multiphysics technology.

Time & space aware

Time-conditioned prediction captures how physical fields evolve across the simulation domain.

Geometry & material aware

The model represents the shape of the engineering system, the space it occupies and its materials.

Boundary, initial & source aware

Boundary conditions, starting states and physical sources enter the model explicitly.

Multiphysics aware

Coupled physical fields are brought together in one representation, connecting interacting physical processes.

Scale aware

Dimensionless numbers — ratios describing the physical regime — encode changes in scale.

Reference comparisons

Predictions are compared with analytical solutions and numerical simulations for the case being studied.

Selected shock-tube benchmark · Company-reported results

99.8%

Agreement with the analytical solution in the selected test.

21,900×

Acceleration versus the numerical physics calculation in that test.

These figures describe one reported comparison, not every application. View the investor presentation · Explore benchmark evidence

Our vision: World Physics Models

Models that connect coupled physical processes, predict the consequences of actions and learn from verification. TUNet-AI is part of our research pathway toward this broader physical intelligence.

Explore TUNet-AI and the engineering demonstrations →
METAL-HYDRIDE HYDROGEN COMPRESSION

See the real four-pipe system and its digital twin

Explore the compressor assembly and see how its pipes exchange heat with a metal-hydride bed. The interactive model explains the system and its simulation limits.

Current physical hardware: Hystore Tech Limited, Participant Identification Code (PIC) 900935225

Historical prototype provenance: HYSTORE TECHNOLOGIES LIMITED, a separate legal entity associated with ATLAS-H2 and ATLAS-MHC

Digital twin and visualisation: KYAMOS

Open the compressor page and full-size images →
Rendered cutaway of a concentric metal-hydride pipe with the hot and cold water annulus, steel wall and porous hydride bed labelled.Concentric hydride pipe cutawayRendered four-pipe metal-hydride compressor system with gauges, valves and a labelled pipe cutaway inset.Four-pipe compressor system

Engineering Software for the CAE Industry

KYAMOS LIMITED develops advanced multiphysics and AI-powered engineering software for scientists, engineers, researchers, and industrial innovators. The company provides Computer Aided Engineering solutions through in-house algorithms, high-performance GPU computing, and cloud-based simulation workflows.

Founded through the IDEA Incubator/Accelerator Programme of the Bank of Cyprus, KYAMOS combines scientific research, software engineering, and high-performance computing for demanding engineering applications.

Our Services

Multiphysics Simulations: Coupled engineering systems involving flow, electromagnetics, thermal behaviour, semiconductors, and energy systems.

Cloud Computing: Cluster-based GPU workflows for scalable research and industrial simulation.

Green Technologies: Wind energy, batteries, thermal systems, inverters, smart grids, and sustainable engineering applications.

Data Science & AI: Surrogate modelling and intelligent automation workflows for faster engineering decisions.

Core Capabilities

Four integrated capabilities — from raw GPU compute to intelligent surrogate models.

Agentic AI

Agentic AI Services

AI assistants that help prepare models, organise engineering tasks and explain results, with checks built into the workflow.

GPU Infrastructure

GPU Infrastructure

80 graphics-processor dies across 10 compute servers, connected for parallel simulation and AI workloads. Capacity is company-stated; project availability is confirmed separately.

Multiphysics Models

Multiphysics Models

Research tools for radio-frequency systems, electromagnetics, semiconductors, wind energy, batteries and heat transfer. Each model has its own evidence and limits.

Surrogate AI

Surrogate AI

AI prediction models trained on simulation results for fast design exploration. Check the demonstrated test cases, error definitions and limits before applying a model.

Funded Research

€2.3M across 14 competitive HPC projects.

Flagship Project: REALISATION-GREEN-CLOUD

REALISATION-GREEN-CLOUD develops cloud-based ICT multiphysics software for green technology industries, targeting customized simulation modules for RF filters, electromagnetics, semiconductors, wind turbines, batteries, and thermal applications.

IDEA

completed

IDEA Incubator/Accelerator Program

Bank of Cyprus

€62,000·OTHER
Jul 2018→Jan 2019

REALISATION-CLOUD

completed

KYAMOS REAl Live SimulATION services - CLOUD computing on GPU

START-UPS/0618/0058

€89,962·RIF
May 2019→Jan 2021

REALISATION-LBM

completed

Lattice Boltzmann GPU REAl Live SimulATION

CONCEPT/0618/0012

€41,000·RIF
Dec 2019→Nov 2020

REALISATION-COVID19

completed

GPU-REAl LIve SimulATION-COVID19

CONCEPT-COVID/0420/0025

€35,700·RIF
Jan 2021→Oct 2021
View all 14 funded projects

Industries

Engineering sectors where KYAMOS simulation, GPU computing, and AI surrogate workflows are applied.

View All Industries
Energy

Energy

Explore heat storage, batteries, renewable energy and electrical-grid studies.

Semiconductors

Semiconductors

Study device heat transfer and charge-carrier transport.

Environmental

Environmental

Explore oil-spill transport, fluid flow and environmental modelling.