Rockwell Automation Joins Anthropic's Project Glasswing, Putting Claude Mythos 5 to Work on Industrial Cyber Defense

The Partnership

Rockwell Automation, the largest company dedicated to industrial automation, announced on September 15, 2026 that it has joined Anthropic's Project Glasswing. The program gives approved organizations controlled access to Claude Mythos 5 for defensive security work.

Rockwell's stated aim is to apply the model to the software and connected systems that run on plant floors, accelerating how quickly its teams find and close security gaps. The activities named in the announcement are discovery, validation, prioritization, and remediation of vulnerabilities.

Tony Baker, Rockwell's Vice President and Chief Product Security Officer, Digital Trust, framed the move around defense. "Rockwell Automation plays a central role in helping manufacturers operate safely, securely, and reliably," he said. "By participating in Project Glasswing, we are applying advanced AI capabilities in a controlled, defensive way to help identify and address vulnerabilities faster and support the resilience of the industrial ecosystem."

What Project Glasswing Is

Anthropic launched Project Glasswing in April 2026 to put its models in the hands of defenders working on critical software. The program began with about 50 organizations and has grown by roughly 150 more, reaching more than 15 countries.

The organizations span critical-infrastructure sectors: power, water, healthcare, communications, and hardware. Access is controlled and scoped to defensive use, so participants apply the model to hardening their own software and connected systems rather than to open-ended research.

Why It Matters for Manufacturers

Factory networks have gotten more connected every year, and each new link between a controller, a sensor, and a business system widens the attack surface. Rockwell's installed base of programmable controllers, drives, and its FactoryTalk software sits at the center of that surface for a large share of North American plants, which is why a product-security move from Rockwell reaches well beyond its own catalog. The company has been building out operational-technology security since its 2023 acquisition of Verve Industrial.

Vulnerability triage is the bottleneck the AI is aimed at. Security teams routinely face more reported flaws than they can validate and rank by hand, and the gap between disclosure and patch is where operational-technology incidents happen. A model that can help sort real exposure from noise and draft the fix compresses that window, which matters more on a plant floor than in an office, because downtime on a production line carries a direct cost.

What to Watch

The framing here is defensive and controlled. Rockwell is exploring how Claude Mythos 5 fits its security workflow, so the near-term signal is which security practices come out of the program, ahead of any feature customers can buy. For manufacturers running connected lines, the thing to track is whether AI-assisted vulnerability work shortens real patch cycles on operational-technology systems, and whether the approach spreads from the program's roughly 200 organizations into the wider supplier base.

Sources

• Rockwell Automation Joins Anthropic's Project Glasswing to Advance AI-Enabled Cyber Defense for Industrial Critical Infrastructure, automation.com
• Rockwell Automation Joins Anthropic's Project Glasswing, Business Wire
• Rockwell Automation Joins Anthropic's Glasswing AI Project, StockTitan
• Rockwell joins Anthropic's Project Glasswing programme, SecurityBrief

Companies mentioned

• Rockwell Automation
• Anthropic