Palo Alto Networks Launches Prisma AIRS AI Gateway for Enterprise Security

Palo Alto Networks has announced the general availability of its Prisma AIRS AI Gateway, a new offering positioned as the AI control plane for enterprises. This gateway aims to help organizations effectively manage and secure their expanding AI footprint Source. The launch delivers on innovations from the Portkey acquisition, six weeks after its close.

The Challenge of Uncontrolled AI Adoption

Enterprises are increasingly relying on AI for various functions, from coding assistance to customer interactions. However, this rapid adoption often creates a significant gap in oversight and security. Palo Alto Networks noted a dramatic increase in AI activity, with AI transaction volume growing twelve-fold in six months and "MCP activity climbed from 11% late last year to 41.4% by mid-2026" Source. This surge in AI usage brings challenges such as:

• Shadow AI: Unsanctioned AI tools and usage can lead to hidden costs and unknown security risks.
• Data Exposure: Sensitive data can be inadvertently exposed through AI interactions, bypassing existing security policies.
• Agent Overstep: Autonomous agents operating with broad permissions can increase the risk of data breaches and reputational damage due to a lack of clear accountability.

Traditional security measures often force businesses to either block AI traffic entirely or permit it with little governance, neither of which is a sustainable strategy for a modern, AI-driven enterprise.

Accelerating AI Adoption with Control

The Prisma AIRS AI Gateway is designed to provide a single control plane for all AI interactions. It sits inline between AI applications, models, and agents, offering unified enforcement for operational and security controls Source. This allows organizations to:

• Discover AI Usage: Gain visibility into which applications, models, users, and teams are active, what data they access, and associated costs.
• Govern AI Interactions: Enforce central rules for model access, tool use, and budgets, and inspect prompts and responses to prevent data leaks.
• Secure Every Agent: Verify agent identities and apply just-in-time, least-privilege access to ensure autonomous systems only access what is necessary.

Key Capabilities of the AI Gateway

The gateway offers several core capabilities:

• Observability: Provides a unified view to track usage, users, projects, token counts, latency, and costs.
• Governance: Enables central definition of approved models and tools, and tracking of costs and usage for FinOps and budget management.
• Coding Assistant Security: Protects credentials, proprietary code, and development systems by replacing raw provider keys with scoped credentials.
• Operational Controls: Applies data protection, usage limits, and policy checks in real-time, ensuring traffic distribution and outage resilience.
• Agent Identity Security: Establishes verifiable, ephemeral identities for agents, allowing only authenticated agents to make approved calls.
• Runtime Security: Inspects prompts and responses inline using Prisma AIRS AI Runtime Security to prevent data exfiltration and prompt injection attacks, aligning with OWASP LLM Top 10 and OWASP Top 10 for Agentic Applications guidelines Source.

Palo Alto Networks highlights that the Prisma AIRS AI Gateway is built on an architecture that has processed over 68 trillion tokens monthly, with sub-millisecond routing latency and 99.999% availability, ensuring it can handle demanding enterprise AI workloads Source.