Mindgard Closes $30M Series A as AI Red-Teaming Emerges as Distinct Procurement Category

The round was led by Album VC with participation from Karma Ventures and prior backers. Mindgard, spun out of Lancaster University in 2022, sells continuous attack-surface mapping and runtime controls aimed at the psycho-technical vectors unique to agents and models. The capital will expand sales and engineering focused on financial services, healthcare, and semiconductor verticals where procurement records already show rising AI pilot spend.

Public evidence includes documented code-execution flaws in Cursor and defects in Google Antigravity and ChatGPT. These findings align with parallel disclosures from other red-team tools and indicate that model supply chains remain exposed to prompt-injection and agentic escalation paths that traditional application security does not cover. Funding patterns mirror recent raises at Corma and Oligo, both targeting runtime AI defense.

The concentration of capital in AI-specific red-teaming suggests buyers are shifting from generic model-risk frameworks toward operationalized offensive testing. Contract vehicles and job postings at defense-adjacent primes now list AI attack-surface requirements, indicating procurement is moving ahead of standards bodies. Expect Mindgard deployments to surface additional zero-days in production agents within regulated sectors by mid-2025.