AI startup Virtue AI launches with $30M to help enterprises deploy generative AI without security risks

Virtue AI, a new AI startup focused on AI safety and security, just came out of stealth with $30 million in combined Seed and Series A funding. The round was led by Lightspeed Venture Partners and Walden Catalyst Ventures, with participation from Prosperity7 and returning backers like Factory, Osage University Partners, Lip-Bu Tan, Amarjit Gill, and Stanford professor Chris Re.

The company is built around a straightforward mission: help businesses adopt generative AI without the usual security headaches. Too often, companies are forced to choose between shipping fast and staying safe. Virtue AI wants to end that tradeoff.

Virtue AI was founded by Bo Li, Dawn Song, Carlos Guestrin, and Sanmi Koyejo—a team of respected academics with decades of combined experience in AI research, with a particular focus on safety and security. Over the years, they witnessed a troubling pattern across the industry: enterprises either rushed AI products to market and accepted the risks, or bogged down innovation with outdated, manual safeguards.

After seeing too many preventable security failures in real-world deployments, they set out to change that—building Virtue AI to turn their deep expertise into practical tools that help companies move fast without compromising safety. That gap is what led to Virtue AI.

The startup is betting big on the idea that traditional cybersecurity approaches don’t cut it for AI. As companies begin to plug generative models into real products and workflows, the threats start stacking up—from prompt injection attacks to data leaks and jailbreak attempts. And as AI becomes more central to how companies operate, regulators are watching closely.

“We saw companies struggling with the same challenges repeatedly—subpar evaluation methods, inefficient guardrails, and manual processes that created bottlenecks in AI deployment pipelines,” said Bo Li, co-founder and CEO of Virtue AI. “Our team has dedicated decades to solving these exact problems. Virtue AI transforms that expertise into practical solutions that eliminate the false choice between innovation and safety.”

At the heart of Virtue AI’s platform are three key tools built to address the most pressing security challenges enterprises face with generative AI.

The first is VirtueRed, an automated red teaming system that handles testing across more than 320 risk categories—including hallucinations, privacy leaks, jailbreaks, and prompt injections. By shifting away from manual testing, VirtueRed significantly reduces both the cost and time required to assess model vulnerabilities.

Then there’s VirtueGuard, a set of high-performance guardrail models that operate across text, images, audio, video, and code—in over 90 languages. The company says these models aren’t just fast—they’re 30 times faster than existing solutions and deliver up to 50% better performance when it comes to filtering risky or unwanted outputs.

Rounding out the stack is VirtueAgent, an AI security agent built to understand an organization’s internal security policies and external regulatory obligations. It automates what are often tedious, manual processes, helping companies meet compliance requirements without slowing their development cycles.

Together, these tools form the foundation of Virtue AI’s mission: helping businesses deploy generative AI responsibly, without having to compromise on speed or safety.

The company’s tech is grounded in academic research that’s been recognized by the NSA, including a widely cited paper that helped expose core risks in foundation models.

It hasn’t taken long to gain traction. Since launching, Virtue AI says it has already secured several major enterprise clients.

One of them is Uber. “Uber leverages Generative AI to deliver magical experiences for our end users,” said Kai Wang, Group Product Manager at Uber’s AI Platform. “We’ve been collaborating closely with Virtue AI to implement robust content safety guardrails for our Gen AI applications, ensuring they are safe, responsible, and aligned with our community standards.”

Glean CEO Arvind Jain echoed that sentiment. “At Glean, we’re committed to building AI that organizations can trust—secure, reliable, and enterprise-ready. Every company has unique security requirements, and we strongly believe in providing the flexibility to choose the approach that best aligns with those needs. Our collaboration with Virtue AI helps us stay ahead of emerging threats and deliver on our promise to keep users in control and their data protected.”

Investors are clearly buying in.

“Virtue AI is shaping the future of GenAI security,” said Lip-Bu Tan, Founding Managing Partner of Walden Catalyst Ventures. “Combining foundational research with advanced algorithms, Virtue AI is tackling the most critical vulnerabilities in AI systems head-on. Their disruptive technology and impressive market traction demonstrate the potential to redefine how organizations across the globe approach AI safety and security.”

Lightspeed’s Guru Chahal and James Alcorn added, “Built on the fundamental AI safety research of its cofounders, Virtue AI has quickly established a leadership position atop the fast-growing AI red-teaming and guardrails category. We are thrilled to partner with the team to help them capitalize on a crucial industry inflection moment, as enterprises demand increasingly robust and sophisticated safety solutions for production LLM deployments.”

The company plans to use the funding to build out its platform, grow its customer base, and bring in more AI safety talent.

For a space filled with hype, Virtue AI is focusing on what could be one of the most critical pieces of the puzzle: keeping AI secure, usable, and compliant without slowing teams down.