Palo Alto buys Portkey to secure AI agents

Palo Alto Networks announced on April 30, 2026 that it intends to acquire Portkey, a provider of AI Gateways. The plan is to make Portkey the AI Gateway layer inside Prisma AIRS, Palo Alto’s platform for securing AI applications and autonomous agents. The deal is expected to close in Palo Alto Networks’ fourth fiscal quarter of 2026, subject to customary closing conditions.

The original source is Palo Alto Networks’ press release, “Palo Alto Networks to Acquire Portkey to Secure the Rise of AI Agents,” plus the company’s own blog post on a unified AI Gateway for agent governance.

This is more than another cybersecurity acquisition. It shows how quickly the AI agent is moving from experiment to governance problem. When agents get access to internal systems, MCP servers, APIs and data, the traffic between model, tool and agent becomes a new control surface. Palo Alto describes Portkey as a central layer for monitoring, routing and securing AI transactions, with policy enforcement, audit logs, semantic routing, failover, quotas and access to more than 3,000 LLMs and MCP tools through one interface.

For CIOs and CISOs, the lesson is straightforward: the agent platform needs the same discipline as identity, networking and cloud. It is no longer enough to ask which model the company uses. The questions become: Who owns agent traffic? Where are decisions logged? Which policies stop an agent in real time? How are cost and data leakage contained when agents start talking to each other?

Portkey says it already processes trillions of tokens per month. If Palo Alto can integrate that layer into Prisma AIRS, the AI Gateway category becomes harder to dismiss as a developer tool. It becomes security architecture.