Polaris Inference closed a £6.2 million Series A round on 27 June, led by Edinburgh-headquartered Techscot Ventures with participation from the Scottish National Investment Bank. The company, operating out of a converted warehouse on Constitution Street in Leith, has spent three years building software that lets hospitals and local authorities run artificial intelligence models directly on low-power hardware — no cloud connection required. This month, that quiet work became very loud, very fast.
The timing matters. Europe is having a difficult summer: heatwaves have killed thousands across France, energy grids are under pressure, and governments from Warsaw to Edinburgh are being pressed to modernise critical infrastructure on tight budgets. Running AI workloads in the cloud costs money and introduces latency. Polaris's pitch is blunt — its software, called EdgeCore, can cut inference costs by up to 60 percent by pushing computation onto chips already installed in devices, whether that's a medical scanner at the Royal Infirmary or a traffic sensor on Princes Street.
What EdgeCore Actually Does — and Who Is Already Using It
NHS Lothian signed a pilot agreement with Polaris in March 2026, deploying EdgeCore across eight radiological units at the Royal Infirmary of Edinburgh on Little France Crescent. The software runs diagnostic-support models on existing Nvidia embedded modules inside the scanners themselves rather than routing patient imaging data to a remote server. The pilot covers roughly 1,400 scans per week. NHS Lothian has not published results yet, but the deal gave Polaris enough credibility to attract the Series A.
City of Edinburgh Council is also testing the technology. A six-month trial begun in January is using EdgeCore on 34 air-quality monitoring nodes across Leith Walk and the Old Town, running predictive models that flag pollution spikes before they breach legal thresholds under the UK Environment Act 2021. The council pays no per-query cloud fee under the arrangement — a detail that caught the attention of procurement officers in Glasgow and Dundee, according to people familiar with the discussions.
Polaris employs 41 people, the majority hired from the University of Edinburgh's School of Informatics on Crichton Street and from Heriot-Watt University's robotics programme in Riccarton. The company's chief technology officer previously led embedded systems work at a well-known semiconductor firm in Cambridge before relocating north in 2023. Staff have access to desk space at CodeBase on Argyle House on Lady Lawson Street — Edinburgh's largest tech incubator — where Polaris first incorporated before outgrowing a shared office in 2024.
The Competitive Picture and What Comes Next
Edge AI is not a new concept, but the market is fragmenting fast. Arm, Intel and a cluster of Israeli and Taiwanese chipmakers are all pushing hardware-level solutions. Polaris is not competing on silicon — it is betting that most organisations already own adequate hardware and simply need better software to unlock it. That is a narrower, more defensible niche than trying to unseat established chip architectures.
The £6.2 million round will fund a sales office in London, planned for September 2026, and a formal product launch at the Web Summit in Lisbon in November. The company has also applied for a place on the Scottish Government's CivTech programme, which matches startups with public-sector challenges and has previously supported 47 companies since its 2016 launch.
For anyone working in health tech, local government procurement or infrastructure planning in Edinburgh, the practical step is straightforward: request a copy of Polaris's white paper on NHS Lothian deployment costs, published on their site on 1 July, before the London office opens and the pitch gets polished for a southern audience. The ground-floor terms tend to come from the Leith warehouse, not the Westminster satellite.