Particle.news

Mindgard Raises $30 Million to Scale Autonomous AI Red‑Teaming

The funding will let the company expand teams and commercial deployments for its continuous platform that maps and tests AI models and agents for runtime exploits.

Overview

  • Mindgard closed a $30 million Series A led by Album VC with participation from Karma Ventures and returning investors .406 Ventures, Atlantic Bridge, IQ Capital, and Lakestar.
  • The company’s core product runs automated red‑teaming that maps an AI attack surface, simulates adversarial chains and jailbreaks, and provides runtime protections for deployed models and agent workflows.
  • Mindgard says its platform has found more than 150 vulnerabilities, including a Cursor IDE zero‑day that allowed code execution, data‑leak defects in Google Antigravity, and guardrail bypasses in ChatGPT’s image generation.
  • Proceeds will fund hires in product, engineering, sales, and marketing and accelerate rollouts into finance, gaming, healthcare, pharma, semiconductors, and other regulated sectors.
  • The raise comes as demand grows for AI‑specific security tied to frameworks such as MITRE ATLAS, OWASP, and Gartner’s AI TRiSM and after a wave of consolidation in the AI security market that is pushing enterprises to embed runtime defenses.