Lancaster University spinout Mindgard lands $30M after breaking Cursor, ChatGPT, and Google’s AI guardrails

Mindgard has closed a $30 million Series A, pushing its total funding to almost $42 million. The round was led by Album VC, with Karma Ventures joining as a new backer and .406 Ventures, Atlantic Bridge, IQ Capital, and Lakestar coming back for another turn.

The pitch is simple enough: don’t wait for AI systems to fail in production, go looking for the cracks now. Mindgard says its platform keeps up automated red-teaming against models, agents, and applications, hunting for shadow AI, testing systems for vulnerabilities, and watching them after launch. The company likes proof, not promises, and says its database now includes more than 150 disclosed vulnerabilities in real products.

That proof includes some loud names. Mindgard says its researchers found a zero-day in Cursor that let its AI editor run any code without phishing links or malware. The same team also leaked Google Antigravity data across trust boundaries and bypassed ChatGPT’s image-generation guardrails. Mindgard disclosed all three issues.

The company started in 2022 in a Lancaster University lab built on more than ten years of research by Dr. Peter Garraghan, who is now chief science officer. James Brear, who previously led Swimlane, Veriflow, and Procera, became CEO in October 2025 as the business shifted from research toward enterprise sales. Mindgard says it is already working with customers in financial services, pharmaceuticals, gaming, digital services, semiconductors, and healthcare.

This is happening inside a market that is getting crowded and expensive fast. Lakera and Protect AI have already been acquired, and Mindgard’s backers clearly think the buyers are still out there. The new money is meant for product, engineering, sales, and marketing, which is usually code for one thing: turn the clever demos into contracts before someone larger decides the whole category belongs inside a bigger security platform.