Virtue AI is an AI-native security and compliance platform that helps enterprises secure their AI systems and agents against threats like prompt injection, hallucinations, and data poisoning. It was founded in 2024 by leading AI safety researchers Bo Li, Dawn Song, Carlos Guestrin, and Sanmi Koyejo.
Virtue AI is an AI-native security and compliance platform that helps enterprises secure their AI systems and agents against threats like prompt injection, hallucinations, and data poisoning. It was founded in 2024 by leading AI safety researchers Bo Li, Dawn Song, Carlos Guestrin, and Sanmi Koyejo.
People
Updated 05/18/26CEO
Founder & Chief Scientist
Founder & Board Director
Founder & Chief Responsible AI Officer
GTM, Strategic International
Pre-Seed Investor
Investor
Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- $30,000,000
Org Details
Updated 05/18/26Virtue AI is an AI-native security and compliance company founded in April 2024 and headquartered in San Francisco, California. The company was co-founded by four prominent AI researchers: Bo Li (CEO), Dawn Song (Board Director), Carlos Guestrin (Chief Scientist), and Sanmi Koyejo (Chief Responsible AI Officer), who together bring decades of foundational AI safety and security research from UC Berkeley, Stanford, UIUC, and other leading institutions. The company's platform is designed to fill critical security gaps that arise when enterprises deploy generative AI. Traditional cybersecurity tools do not address AI-specific threats such as prompt injections, jailbreaking, hallucinations, data poisoning, and model extraction. Virtue AI's products are built specifically for these risks. The platform consists of four main components. VirtueRed provides continuous, algorithmic red-teaming across more than 1,000 risk categories in a scalable, repeatable, and cost-efficient manner. VirtueGuard delivers real-time, multimodal guardrails with sub-10ms latency, covering text, image, audio, video, and code across more than 100 languages. VirtueGov provides unified governance for AI agents, models, and applications. AgentSuite offers integrated protection across the full AI agent lifecycle, including protection at the prompt, action, network, and MCP tool levels. Virtue AI integrates with major AI platforms including Claude SDK, OpenAI, LangChain, Google ADK, Amazon Bedrock, and GitHub Copilot, as well as enterprise tools such as Salesforce and ServiceNow. The company has been named a representative vendor in Gartner's Market Guide for Guardian Agents. Clients span finance, healthcare, and enterprise software, with notable customers including NVIDIA, Anthropic, Uber, and Glean. In April 2025, Virtue AI announced $30 million in combined Seed and Series A funding, co-led by Lightspeed Venture Partners and Walden Catalyst Ventures, with participation from Prosperity7 and other investors including Factory, Osage University Partners, Lip-Bu Tan, Amarjit Gill, and Chris Re.
Theory of Change
Updated 05/18/26Virtue AI believes that AI systems deployed at enterprise scale are highly vulnerable to adversarial attacks, misuse, and policy violations that existing cybersecurity frameworks cannot address. By providing automated red-teaming, real-time guardrails, and governance tooling, they aim to make safe AI deployment accessible and scalable for any organization. Their implicit theory of change is that broad adoption of robust AI security infrastructure reduces the incidence of AI-related harms — including misuse by bad actors, unintended harmful outputs, and compliance failures — thereby building public trust in AI and reducing the probability that unsafe AI deployments cause large-scale negative outcomes.
Grants Received
Updated 05/18/26Projects– no linked projects
Updated 05/18/26Discussion
No comments yet. Be the first to share your thoughts.