Gray Swan
A Carnegie Mellon-born Pittsburgh AI red-teaming and safety vendor used directly by frontier AI labs.
Visit Website ↗ + Add to Compare Claim This CompanyOverview
Gray Swan, founded in 2024 and based in Pittsburgh, gives security and AI-safety teams tools to assess how AI models and agents can behave in unintended ways and then provides controls to prevent that behavior. It grew out of Carnegie Mellon’s AI safety research team and is used directly by frontier AI labs to red-team their own models.
It has raised a Series A and carries unusually deep academic and technical credibility in AI red-teaming.
Innovation Matrix Assessment
Progressed from $5M early-stage rounds in 2024 and 2025 to a $40M Series A, steadily building out frontier-lab-grade AI red-teaming tooling.
Reported direct engagement with major frontier AI labs suggests meaningful operational credibility, though customer specifics are undisclosed.
Raised a $40M Series A in May 2026 co-led by Wing Venture Capital and Madrona, with Snowflake Ventures and Hudson River Trading participating.
AI model/agent red-teaming and safety assessment born directly out of academic AI safety research is a differentiated, credible approach.
Direct adoption by frontier AI labs to test their own models is meaningful third-party validation, even though not independently itemized.
Assessing and preventing unintended AI model/agent behavior is an urgent, durable need as frontier AI capabilities and deployment scale rapidly.
Why CISOs Should Care
Gray Swan gives CISOs and AI safety teams tools to assess how their AI models and agents can behave in unintended ways, then provides the security controls to prevent that behavior -- used directly by major frontier AI labs.
What Makes It Different
Gray Swan grew out of Carnegie Mellon's AI safety research team and is trusted directly by frontier AI labs to red-team their own models, giving it unusually deep academic and technical credibility in AI red-teaming specifically.
The Matrix Verdict
53/100 — INCREMENTAL INNOVATOR
A highly credible, CMU-born AI red-teaming and safety vendor trusted by frontier AI labs themselves -- a strong signal of genuine technical differentiation in a crowded AI security field.
Editorial Note: Claims vs. Verified Findings
The claim of being 'trusted by every major frontier lab' is a company-stated positioning claim from press materials, not independently itemized against each lab.
Sources
Alternatives to Gray Swan
Adaptive Security
AI-driven platform that simulates deepfake, voice, and multichannel social-engineering attacks to train and test organizations against next-generation phishing.
Quilr
CISO ReviewedEarly-stage agentic AI security startup building a 'Service-as-Software' platform to guard against human-related breaches and secure AI agent…
Zenity
Governance and security platform for AI agents and low-code/no-code development, securing agent identity, permissions and behavior across the…
Gray Swan AI
Carnegie Mellon-linked startup building automated red-teaming and runtime guardrails that frontier AI labs use to stress-test their models…
Tenzai
An agentic AI penetration testing startup building autonomous 'AI hackers' to find and validate exploitable vulnerabilities at a…
Reco
Reco secures the "agentic ecosystem" — mapping what AI agents can access across SaaS and enterprise apps, detecting…