Updated on 5 October 2026
THYEN – Two-Hemisphere Physical AI for Verified Embodied Intelligence & Energy-Efficient AI Hardware
Founder & Inventor – THYEN Physical AI at HyperCell Energy Ltd
London, United Kingdom
About
THYEN is a proprietary UK Physical AI and embodied-intelligence architecture being developed by HyperCell Energy Ltd.
The problem: As AI moves from software into robots and autonomous physical systems, AI-generated decisions can produce real-world consequences. THYEN is designed so that model-generated intentions do not automatically become physical actions.
The architecture: THYEN uses a two-hemisphere approach combining reasoning, independent verification, safety/authority control and physical execution. A separate execution layer controls authorised actions before they reach physical actuators.
What exists today: A working heterogeneous local development environment incorporating NVIDIA Jetson edge computing, local AI inference, multiple AI-model roles, sensing, voice interaction and controlled actuator interfaces. A physical robotic demonstrator tests the complete chain from natural-language instruction through verification and safety control to authorised physical action and sensory feedback.
Next stage: Independent benchmarking, expanded embodied-AI validation, energy-per-task measurement, a more capable robotic platform, distributed compute optimisation and progression towards FPGA/SoC and specialised energy-efficient Physical AI hardware, with potential future heterogeneous/chiplet and neuromorphic implementation where supported by evidence.
Investment: HyperCell Energy Ltd is seeking £600k–£750k strategic pre-seed investment to finance the next integrated engineering and validation programme.
We are particularly interested in investors with expertise in Physical AI, robotics, semiconductors, edge AI, AI safety or energy-efficient computing and potential participation in the Innovate UK ACT & Semiconductor Investor Partnership programme.
THYEN forms part of a wider protected technology portfolio. Detailed proprietary architecture is available only through an appropriate confidential due-diligence proces.
Organisation
Similar opportunities
Investment
Finchetto – Heterogeneous Photonic Integration for AI-Scale Optical Switching
Mark Rushworth
CEO at Finchetto
Guildford, United Kingdom
Investment
Knowing Without Seeing (KWS) – Semiconductor Sensing + Edge AI for Human-State Understanding
Arun Luther
Founder & CEO at Genesis Polymer Innovations Ltd
Birmingham, United Kingdom
Investment
Evrmind: hardware-enabled verification for AI, £3.5m pre-seed, seeking lead investor
Ibrahim Vandenberg
CEO & Co-founder, Evrmind at Evrmind Limited
London, United Kingdom