| MatNex (Materials Nexus)Companies · also on Labs | Machine-learning and quantum-mechanics software that predicts how new material compositions will perform, to cut lab experiments.Profile- Legal name
- Materials Nexus Ltd (per MatNex/Viridien release)
- Founded
- Not stated (Cambridge spin-out)
- Founders
- Jonathan Bean — CEO and co-founder; Robert Forrest — CTO and co-founder
- Chief executive
- Jonathan Bean
- Chief scientist or research lead
- Robert Forrest — CTO
- Other leaders
- Richard Tomsett — Machine Learning; David Gunn — Scientific Computing; Freddie Markanday — Platform Science; Carys Allsopp — Product Management; Kristy Agostini — Finance; Georgie Riddle — Operations
- Hq
- Cambridge (England: East of England) per UKTN 2023; current site states no HQ; other_offices: not stated
- Headcount
- Not stated (team page lists 25 people, Oct 2026)
- What it does
- MatNex designs new materials to a customer's specification with AI and physics simulation, then tests them and patents them. Its first focus is magnets that use fewer rare-earth elements, for electric vehicles, wind turbines and data storage; it plans to move into batteries, semiconductors and superconductors.
- Research focus
- "The critical materials design company"; machine-learned interatomic potentials for magnetic materials; generative design for superconductors.
- Models and products
- MatNex design engine — first-principles computation plus machine learning — not stated; mMACE (magnetic MACE) — interatomic potential for magnetic materials — 9 Apr 2026; rare-earth-free permanent magnet — AI-designed magnet — reported 19 Jun 2024
- Customers or partners named
- Viridien — partnership to scale computing capacity (AI, HPC) — 20 Feb 2025 — https://matnex.ai/news/matnex-and-viridien-partner-to-accelerate-ai-powered-materials-discovery; UKAEA — contract for project SHINE — 2 Feb 2026 — https://matnex.ai/news
- Selected papers
- Equivariant Many-body Message Passing Interatomic Potentials for Magnetic Materials (Ho, …, Tomsett, Forrest, Bean, Csányi, Ortner) — arXiv — 9 Apr 2026 — https://arxiv.org/abs/2604.08143; Reinforcement learning expands generative design for superconductors — OpenReview — 26 Apr 2026 — https://openreview.net/pdf?id=Ac24KyyfyP
- Rounds found
- 26 Jul 2023 — seed — £2m — Ada Ventures — High-Tech Gründerfonds, University of Cambridge, MD One Ventures, Michael Eisenberg, angels (already DEAL-LB-090)
- Awards and programmes
- KPMG Black Entrepreneurship Awards winner — Oct 2023; UKAEA contract (project SHINE) — Feb 2026
- Status now
- Trading as MatNex; holds a UKAEA contract (project SHINE) from February 2026 and published magnetic-materials modelling work in April 2026.
Labs Round 8 deep profile (31 Oct 2026), as the researchers wrote it; confidence: verified (2023 seed: trade plus company news list); trading name and UKAEA contract one source. Sources - matnex.ai/news/matnex-and-viridien-partner-to-ac
- matnex.ai/news
- matnex.ai/
- uktech.news/ai/materials-nexus-2m-funding-202307
- arxiv.org/abs/2604.08143
| Cambridge (CB1 2LA) | Companies House · uktech.news/ai/materials-nexus-2m-funding-202307 · find-and-update.company-information.service.gov. · find-and-update.company-information.service.gov. · assets.publishing.service.gov.uk/media/64bf8d179 | verified |
| PolaronCompanies · also on Labs | London start-up that trains AI on microscope images and material measurements to predict how manufactured materials will behave.Profile- Legal name
- Polaron Ltd (site footer)
- Founded
- Not stated (spun out of Imperial College London "after seven years of research")
- Founders
- Isaac Squires — CEO and co-founder; Steve Kench — CTO and co-founder; Sam Cooper — Chief Scientist and co-founder
- Chief executive
- Isaac Squires
- Chief scientist or research lead
- Sam Cooper — Chief Scientist
- Other leaders
- Steve Kench — CTO
- Hq
- London (England: London), Senna Building, Gorsuch Place, E2 8JF; other_offices: not stated
- What it does
- Polaron builds generative AI that reads microscope images of a material's internal structure and links how a material is made to how it performs, so manufacturers can design and qualify materials such as battery electrodes faster.
- Research focus
- "We are accelerating the design of advanced materials with generative AI"; microstructural image data to link process and performance.
- Models and products
- Polaron platform — characterisation layer and generative design layer; applications in R&D, quality and qualification, modelling and simulation — not stated
- Customers or partners named
- Not stated (release says engineers at manufacturers incl. EV makers have used it; none named)
- Selected papers
- Not stated (site cites Nature Machine Intelligence and Advanced Materials without titles)
- Rounds found
- 3 Feb 2026 — round (stage not named) — $8m — Racine2 (fund led by Serena and Makesense) — Speedinvest, Futurepresent, angels (already DEAL-LAB-2011 / DEAL-SU-IND-042)
- Awards and programmes
- Manchester Prize for AI in science and engineering — £1m winner — date not stated
- Status now
- Trading; raised $8m in February 2026 to expand engineering and roll out generative design tools.
Labs Round 8 deep profile (31 Oct 2026), as the researchers wrote it; confidence: one source. Sources - polaron.ai/about-us
- polaron.ai/newsroom/polaron-raises-8m-to-build-t
| London | Companies House · uktech.news/ai/ai-materials-startup-polaron-rais · find-and-update.company-information.service.gov. · find-and-update.company-information.service.gov. · tech.eu/2026/02/03/polaron-completes-8m-funding- | verified |