Anthropic debuts Claude Ultra with private-cloud inference for regulated industries

AI · 6 min read

Anthropic today unveiled Claude Ultra, a new model variant engineered specifically for private-cloud and on-premise deployments that require strict data residency and auditability. Claude Ultra supports encrypted model weights and inference auditing hooks out of the box, enabling banks, healthcare providers, and government customers to run generative AI without sending prompt data to public inference endpoints.

Alongside the product launch, Anthropic closed a $250 million strategic extension round led by a consortium of sovereign wealth funds and enterprise cloud providers. The financing is earmarked for compliance tooling, secure orchestration layers, and partner integrations with managed service providers focused on regulated verticals.

Anthropic says Claude Ultra will be available in pilot form this quarter and generally available in Q4, with connectors for major private-cloud platforms and a managed offering from select cloud partners. Early customers note that the combination of model accuracy and enterprise controls could shorten procurement cycles for high-risk industries.