Anthropic is launching a cybersecurity initiative for critical infrastructure. The company told Axios about the plan on Thursday.

The Critical Infrastructure Defense Program will bring its most powerful AI models and engineers to companies that protect power grids, water utilities and other critical systems.

The goal is to help operators stop cyberattacks. AI has made attacks easier and cheaper to carry out, putting essential services at greater risk.

The announcement comes as AI-powered hacking grows. This week, CrowdStrike said recent attacks on South Korean banks may have been carried out by a hacker using Chinese and US AI tools, including Anthropic’s Claude Code.

Governments are paying close attention. Officials in the US and elsewhere worry that advanced AI could help attackers breach the systems that keep lights on and water running.

The program will pair Anthropic’s engineers with utility companies and grid operators. The aim is to use AI defensively, to spot and stop attacks faster.

Anthropic CEO Peter Thiel has argued that AI development should continue at full speed. The new program is his answer to critics who say powerful AI mainly helps attackers.