Radware adds MCP server protection to its agentic AI security suite

Radware has launched a purpose-built solution to secure Model Context Protocol servers as agentic AI deployments expand enterprise attack surfaces.

By The Datatech Times Editorial Team AI level 2 of 5
AI level 2 of 5: drafted by our AI editorial assistant from the company's announcement; fact-checked, edited and signed off by Mark Walker, Editorial Director. What the levels mean
A liquid-cooled server rack in a data center displays transparent tubing, circuit boards with blue and green LED lights, under bright artificial lighting.

Radware (NASDAQ: RDWR) has launched Radware Model Context Protocol Server Protection, a security product designed to guard the integration layer between AI agents and enterprise systems. The announcement, made on 14 October 2026, extends the company's existing AI security portfolio, which already includes an Agentic AI Protection module and an LLM Firewall, to cover the MCP servers that connect agents to business applications, data and workflows.

MCP is an open protocol that lets AI agents discover and invoke tools and services within enterprise environments. Adoption is accelerating: a 2026 survey by Stacklok found that nearly half of software companies questioned were already running MCP in production, with 19 per cent reporting broad production use. Radware's argument is that this rapid uptake has created an under-protected attack surface that conventional web application firewalls and DDoS controls were not designed to address.

What the product does

The new offering provides four main capability areas. First, it maps an organisation's MCP infrastructure, inventorying deployed servers, exposed tools, agent identities and communication flows. Second, it applies MCP-aware security controls targeting protocol abuse, authentication attacks, tool enumeration, unauthorised interactions and AI-driven denial-of-service flooding. Third, it includes a governance layer to evaluate exposure, surface weaknesses and enforce policy as deployments scale. Fourth, it monitors agent behaviour continuously, flagging anomalous tool invocation patterns and unusual resource consumption.

The product is available either as a cloud-based service through Radware's Cloud Application Protection Service or as a cloud-augmented add-on to the company's on-premises Alteon and DefensePro platforms, covering public cloud, private cloud, Kubernetes and hardware data-centre environments. Radware did not disclose pricing or name any early-access customers in its release.

Chief technology officer David Aviv said organisations "need to know which MCP servers they have, what those servers expose, which agents are accessing them and whether those interactions are safe," adding that Radware is extending its AI security approach to protect the MCP layer that connects agents to real business systems.

Market context and competitive landscape

MCP server security is an emerging sub-category within the broader agentic AI security market, which itself is still taking shape. Established cybersecurity vendors including Palo Alto Networks, CrowdStrike and Zscaler have each been building out AI workload protection capabilities, while a wave of specialist startups is targeting AI runtime security, prompt injection defence and agent identity management. Radware's move to name MCP servers specifically as a protection target is among the earlier product-level responses to that particular attack vector.

The wider context matters for buyers. Agentic AI systems present distinct security challenges from conventional web applications: agents can chain tool calls autonomously, escalate privileges through misconfigured MCP integrations and exfiltrate data via legitimate-looking API calls. Traditional perimeter controls do not inspect MCP traffic at the protocol level, which is precisely the gap Radware is positioning its new product to close.

Regulatory pressure is also beginning to bear on AI system security. The EU AI Act's obligations for high-risk AI systems include requirements around robustness, accuracy and security logging, and the UK's Cyber Security and Resilience Bill is expected to extend incident reporting requirements to a broader range of digital infrastructure. Organisations deploying agentic AI in regulated sectors will face mounting pressure to demonstrate that the integration layers connecting agents to enterprise data are audited and protected. Radware's governance module appears designed with that compliance need in mind, though the company has not made explicit FedRAMP, SOC 2 or ISO 27001 certification claims for the new product.

Radware has scheduled a webinar for 3 December 2026 to walk through the new offering in detail.

Sector

Topic

Geography