Mindgard raises $30M to handle security for AI models and applications

Mindgard Ltd. has pulled in $30 million in early funding as demand grows for tools that can secure AI models, applications and agents. The Series A was led by Album VC, with Karma Ventures and existing backers joining in alongside .406 Ventures, Atlantic Bridge, IQ Capital and Lakestar.

The company is leaning on a mix of AI security research and offensive security know-how, which is a fairly blunt way of saying it tries to think like an attacker before someone worse does. Founded in 2022, Mindgard began as a spinout from Lancaster University and built on more than a decade of research aimed at stopping attacks on AI systems. It now operates independently from Boston and London.

Mindgard says it has publicly disclosed more than 150 high-impact vulnerabilities in popular AI applications so far. The examples it points to are the sort of thing security teams would rather not discover in production: zero-day code execution in AnySphere’s Cursor, a trusted workplace flaw in Google’s Antigravity, and image-generation guardrail failures in OpenAI ChatGPT.

The funding is meant to help the company expand what it calls its AI Security Platform and meet growing customer demand. Mindgard says the Series A has already accelerated deployment across Fortune 2000 companies and AI-focused firms in financial services, pharmaceuticals, gaming, digital services, semiconductors and healthcare. It also plans to add engineering staff and push harder on product, sales and marketing as AI moves deeper into critical operations.