Palo Alto Networks has announced that Unit 42® Continuous Frontier AI Defense and the Prisma® AIRS™ AI security platform are available on the newly launched OpenAI Marketplace, giving OpenAI’s enterprise customers direct access.
The backdrop is familiar: attackers use AI to find and exploit vulnerabilities at unprecedented speed. According to Palo Alto Networks, breach cycles have shrunk by almost 97%, from weeks to hours, and human-speed defence can no longer keep up with automated, continuous attacks.
An advantage for organisations with OpenAI commitments
Organisations with OpenAI spend commitments can deploy machine-speed, proactive defence while drawing down those commitments, bringing the security investment and the AI budget together.
What is Unit 42 Continuous Frontier AI Defense?
It is an always-on, expert-led, agentic offensive security service. As the environment changes, it finds exposures across applications, identities, cloud infrastructure and network assets, validates whether they are really exploitable and accelerates remediation.
- At its core is a multi-model harness that routes each security task to the best-suited model, combining defender-grade models such as OpenAI’s GPT-5.6-Cyber with other commercial and open-weight models and Unit 42 threat intelligence.
- Non-destructive adversary simulations prove real exploitability and deliver prioritised remediation guidance, including code-level fixes and virtual patches.
- According to Palo Alto Networks, six months of internal testing uncovered two years’ worth of exposures in three weeks. Across more than 75 customer engagements, exposures were found in 100% of assessed organisations, 37% of them high or critical, and over two-thirds of exposures in third-party applications had no known CVE.
Scaling AI safely with Prisma AIRS
As OpenAI customers roll out powerful models, coding assistants and autonomous agents, the attack surface grows: shadow AI, supply chain risk in open-source models, unsafe actions from malicious inputs and agents with excessive permissions.
Prisma AIRS addresses these risks in four areas:
- Visibility: inventory AI apps, coding assistants and agents across endpoints, cloud and enterprise environments, and expose shadow AI.
- Risk identification: scan models and agent artifacts such as skills and MCP servers (AI Supply Chain Security) and red-team apps before deployment.
- Interaction, data and identity security: real-time protection against prompt injection and data exposure (AI Runtime Security), plus verified agent identities and least privilege (Agent Identity Security).
- Central policy and spend: govern connections to LLMs, MCP servers and APIs with the AI Gateway, and track model access, usage and cost across teams.
Palo Alto Networks was recognised as a Market Shaper in the September 2026 Gartner® Emerging Market Quadrant for AI Application Security.
For organisations in Türkiye
The purchasing channel may be the OpenAI Marketplace, but deployment, policy design and integration with existing security infrastructure need local expertise. Albisec supports Prisma AIRS assessments and deployments as your Palo Alto Networks partner. Request a free assessment.
Source: Palo Alto Networks Blog, “Palo Alto Networks Brings Continuous AI Defense and Prisma AIRS to the OpenAI Marketplace”, Munish Khetrapal, 23 October 2026. Images: Palo Alto Networks.
