Anthropic has deployed Claude Mythos into critical infrastructure systems across 15 countries, according to a TechCrunch report. The rollout puts Anthropic's AI into the systems that keep economies running - energy, water, transportation, finance - in multiple nations simultaneously, not just as a productivity tool for employees but integrated into operational workflows.
The commercial logic is straightforward. Critical infrastructure contracts are long-term and hard to exit. Once your AI is woven into how a country manages its power grid operations or financial clearing network, switching costs are enormous. These customers are fundamentally different from a SaaS subscription - closer to embedded infrastructure deals that run for years or decades. That kind of contract durability is valuable in an industry where most revenue is still month-to-month.
The risk side of this is harder to dismiss. Critical infrastructure is exactly where AI errors have cascading consequences. The same confident-but-wrong behavior that makes a chatbot mildly annoying in a productivity context becomes a qualitatively different problem inside a water treatment system or an energy management platform. Anthropic has built Constitutional AI as its core safety approach - a framework designed to make Claude behave reliably and honestly by training it against a set of defined principles - but whether that's sufficient for high-stakes operational contexts is a question regulators in at least 15 countries are now going to be asking.
No public list of which countries are involved has been confirmed as of June 2.