Mind Foundry: applied AI, assurance and industrial research partners

Mind Foundry develops machine-learning software for organisations working with difficult data and consequential decisions. Its current website puts defence and national security at the centre of the offer, with SENTRY and NIGHTINGALE as named products. Its wider business history includes insurance research and industrial collaborations, making the company more specific than a generic provider of artificial intelligence consulting.

The Oxford business grew from university research. Its current leadership page names Lindsay Chadwick as chief executive, Al Bowman as chief commercial officer, and Stephen Roberts and Maike Osborne as co-founders and chief science officers. That current team should be distinguished from the people quoted in older funding and programme announcements.

A private company with new ownership disclosures

Companies House identifies Mind Foundry Limited, number 09882159, as a private limited company incorporated on 20 November 2015. Oxford's account of the business describes the university spinout in 2016. Those dates refer to different parts of its formation.

The current significant-control register provides a more recent ownership reference. Parkwalk Advisors Ltd is recorded with more than 25% but not more than 50% of voting rights. Oxford Science Enterprises Holdings II Limited is recorded with more than 25% but not more than 50% of shares and voting rights. Both notifications are dated 27 January 2026.

Those disclosures describe specified control interests, not a complete shareholder list. In particular, an older strategic investment by an insurance company should not be turned into an assertion that Mind Foundry is wholly owned by that investor. The register and the commercial collaboration history answer different questions.

SENTRY and NIGHTINGALE define the current offer

The current company website emphasises software that can work with existing hardware and support decisions under uncertainty. That is a useful starting point for identifying the product's place in a larger system: Mind Foundry supplies analysis software rather than manufacturing every sensor that produces the data.

SENTRY is marketed as a sensor-fusion product. The company describes integration with third-party applications and different deployment arrangements. Commercially, the relevant proposition is that a customer can derive more value from information produced by several systems without buying all of those systems from Mind Foundry.

NIGHTINGALE addresses acoustic analysis. Its public page describes trials involving Thales, the Ministry of Defence and the Royal Navy, and a separate relationship involving SEA's KraitSense hardware. These are named integration examples, which make the product easier to place in the market than an unqualified claim of general AI capability.

The trial descriptions are the company's account of its work. They establish the partnerships it publicly reports, but do not publish a fleet-wide purchase quantity or a complete revenue figure. For a software supplier, product integration, user evaluation and repeat commercial deployment remain distinct milestones.

Three public routes into defence work

The company has entered defence work through several documented relationships. Each gives a different view of how its expertise reaches a buyer.

On 8 October 2024, DASA announced funding for Mind Foundry under DTEP. BAE Systems was identified as the higher-tier partner for a project involving multiple types of sensor input. The programme announcement described a government grant covering 50% of project value, without publishing a Mind Foundry-specific cash amount.

This is evidence of supported development and industrial collaboration. It does not imply that BAE Systems has acquired Mind Foundry, or that the development grant is a production contract for every potential application. The partner relationship is commercially important because it connects a software specialist with an organisation experienced in delivering larger systems.

A different record comes from Saab UK on 24 January 2024. Saab announced a LookOut Phase 3 contract in collaboration with Mind Foundry, funded through DASA on behalf of the Royal Navy. The stated purpose was to mature and reduce risk in technology for future maritime early warning.

The important distinction is the named contractual lead: Saab announced the award and identified Mind Foundry as its collaborator. This establishes a programme role without assigning Saab's whole contractual responsibility to the software company.

A third example concerns maintenance rather than a sensor product sale. In an October 2024 case study, the Accelerated Capability Environment described a four-week study by Mind Foundry and Frazer-Nash into AI applications for assessment of aircraft wear debris. The study identified potential uses and a brief proof of concept.

That record gives a buyer-side description of discovery work. It is particularly useful because the scope is clear: the organisations explored whether the approach could help, rather than claiming that an entire maintenance process had already been replaced.

Insurance research remains part of the business history

Mind Foundry's industrial relationships extend beyond defence. Oxford's February 2023 announcement identifies the Aioi R&D Lab–Oxford as a joint venture involving Mind Foundry, Aioi Nissay Dowa Insurance and Aioi Nissay Dowa Europe. Its work covered subjects including mobility, insurance and climate.

The relationship also has a financing history. Mind Foundry's 9 November 2020 Series A announcement reported $13.6 million led by Aioi Nissay Dowa Insurance, alongside existing investors. The release described plans to expand use of the platform after an earlier project with the insurer's UK subsidiary. This is evidence of a strategic customer-investor relationship at that date. It is not a substitute for the 2026 register or a measure of present annual sales. Keeping those records separate makes the commercial history more useful.

The joint venture gives the AI company access to a different class of domain problems and research partners. It is not equivalent to a public claim that every insurance operation in the wider Aioi group uses Mind Foundry software.

A September 2024 Oxford announcement describes the laboratory's support for doctoral research and industry-led projects. That provides evidence of a continuing research relationship beyond the original launch.

For business readers, the lesson is specific to Mind Foundry's route to market: its applied research can be organised through a joint venture, a development partner or a customer study. Those arrangements have different commercial economics even when they draw on related scientific expertise.

Revenue and employees from the amended accounts

The amended 2024 annual report reports consolidated turnover of £7,598,964 for the year ended 31 December 2024, compared with restated 2023 turnover of £5,490,978. Note 4 separates £5,252,208 of subscription products, £2,206,262 of one-off services and £140,494 of other turnover.

Note 8 gives an average monthly workforce of 85 employees including directors, compared with 72 in 2023. These are historical reporting figures. The turnover is group-wide and should not be labelled defence revenue; the workforce is an average rather than a current count. The same accounts report an operating loss of £6,129,660.

The split between subscription products and services is particularly useful. It shows that the business cannot be understood only through announcements of research partnerships. There is a recorded product revenue stream alongside separately identified service work.

How to compare Mind Foundry

The Vizgard profile offers a neighbouring view of software supplied into sensing systems. The Fujitsu and St Andrews analysis examines another AI adoption setting where a defined customer problem matters more than a broad capability label.

Mind Foundry's differentiating evidence is the combination of named products, industrial collaborators, university research and reported commercial revenue. The next useful developments would connect that record with repeat customer deployments and a clearer view of which product families drive the business.