Lancaster University spinout Mindgard raises £22m to scale its AI red teaming platform

[DiogoCardoso]

Mindgard, an AI security company spun out of Lancaster University, has raised $30 million, about £22.2 million, in a Series A round to expand its platform for assessing and protecting AI systems, with the round led by Album VC and participation from Karma Ventures alongside existing backers .406 Ventures, Atlantic Bridge, IQ Capital and Lakestar

The company operates dual headquarters in London and Boston, which is a fairly common structure for UK deep tech companies trying to stay close to both British research roots and the larger US enterprise security market where a lot of the biggest AI security buying decisions are actually happening

Mindgard's platform covers shadow AI discovery, continuous red teaming and real time threat protection, aimed at finding vulnerabilities across AI models, autonomous agents and enterprise applications, which positions it less as a single point tool and more as an ongoing security layer meant to sit alongside AI systems as they're deployed rather than just testing them once before launch

The credibility signal that stands out most here is the vulnerability disclosure track record, Mindgard says its platform has been used to publicly disclose more than 150 high impact AI safety and security vulnerabilities in widely used tools, including zero day flaws in Cursor IDE, Google Antigravity, and guardrail bypasses in OpenAI's ChatGPT, which is a genuinely substantial body of real findings rather than just marketing claims about capability

CEO James Brear's framing is worth taking seriously, that AI is creating an entirely new attack surface and that the company is trying to operationalize the expertise of AI security researchers and offensive security practitioners into something enterprises can actually use continuously, rather than positioning the product as simple automated attack tooling, that distinction between packaged expertise and blunt automation is exactly the pitch that tends to attract serious enterprise security budgets

This raise is also a good data point on where UK AI security specifically is heading, a Lancaster University spinout with real disclosed vulnerabilities against major AI products landing a $30 million round suggests investors see continuous AI red teaming as a genuinely distinct and fundable category now, not just a feature bolted onto existing cybersecurity vendors' broader product lines

Lancaster spinout Mindgard raises £22m to scale AI security platform - UKTNwww.uktech.news

[MayanHan]

150 disclosed vulnerabilities including zero days in Cursor and Google Antigravity is a strong track record to back up a Series A raise

[MegaNathan60]

Agreed, thats a lot more credible than just claiming red teaming capability without any public disclosure history to point to

[Always_Reuben87]

Dual headquarters in London and Boston makes sense given how much enterprise AI security buying is concentrated in the US market right now

[NatureBoyRyan65]

Lancaster University spinouts dont get nearly enough attention compared to the usual Oxford and Cambridge deep tech pipeline, good to see one landing a raise this size

[Mia_59]

30 million dollars is a solid Series A for this space, curious what specific expansion plans they have for the new capital beyond vague scaling language

[IronQuarry]

Continuous red teaming as an ongoing service rather than a one time pentest style engagement is clearly where this whole category is heading

[Darren_20]

The disclosed vulnerability count is probably their strongest differentiator right now, thats not easy for competitors to just claim without matching evidence

[SašaJelenič]

Makes sense given how fast AI models and agent behaviors change, a single point in time security test goes stale almost immediately

[KnotKnull]

Shadow AI discovery is the part of their platform that interests me most honestly, most companies genuinely have no idea what AI tools are actually running inside their org

[NeverQuitZach33]

Curious how Mindgard actually differentiates from the growing pile of other AI security and red teaming startups chasing the same enterprise budgets

[Edward71]

Thats becoming a huge blind spot for enterprises fast, employees adopting AI tools without any central visibility is a real and growing risk