Anthropic’s new program will put frontier Claude models, company engineers and threat research in the hands of firms that already defend power grids, water utilities, factories and transportation networks. The Critical Infrastructure Defense Program is the first operational-technology track in Anthropic’s broader Cyber Mission.
Operational technology can be unusually difficult to secure. Industrial controllers and networks are often built to run for decades, while taking them offline for a patch can interrupt essential services. Anthropic says its first step is to work through trusted security, consulting and manufacturing providers rather than directly with every operator.
A small first cohort
The 11 founding partners are Accenture, Booz Allen, CrowdStrike, Deloitte, Dragos, Hitachi, Insane Cyber, Nozomi Networks, Palo Alto Networks, PwC and Rockwell Automation. Anthropic says several of those organizations are already using Claude to find and repair vulnerabilities or help customers do so.
The company has not detailed how the models and engineers will be deployed across each partner. Axios reports that Anthropic has not said whether partners will receive free model access or who will pay the computing costs. The report also notes a central safety question: how fixes can be tested and deployed without disrupting the utilities they are meant to protect.
The broader cyber push
Critical infrastructure is one of two initial tracks in Anthropic’s Cyber Mission. The other is OSS Scanner, a free opt-in service that periodically scans enrolled open-source projects for vulnerabilities and proposes fixes.
Those scanner reports are sent without human review. Anthropic warns that some may include inaccurate findings or incorrect severity ratings, so maintainers will still need to validate what the model produces. The company says it plans to expand the Cyber Mission with more tools, research and resources over time.
Comments