Opaque Systems

AI Security · Confidential Computing · Enterprise AI

Acquisition

Anthropic acquires confidential-AI startup Opaque Systems

The deal shows frontier AI labs now building the data-security layer themselves instead of leaving it to outside partners.

• Anthropic has agreed to acquire Opaque Systems, a startup specializing in confidential computing and AI; financial terms were not disclosed.
• Opaque grew out of a decade-old UC Berkeley RISELab research project and built technology letting organizations run AI on sensitive data with hardware-enforced protections.
• The startup partnered with Nvidia, Intel, Microsoft and Google and had raised a $24 million Series B at a $300 million valuation before the acquisition.
• Opaque recently expanded from confidential computing into 'verifiable AI,' using cryptographic proof of what code ran, where, and under what policies, and had been rolling out open standards for sovereign AI and on-prem deployments adopted in Europe and the UAE.
• Analysts frame the move as evidence that as AI moves into regulated industries, model providers are choosing to absorb the security layer rather than depend on partners for it.
• Owning confidential-AI infrastructure lets Anthropic court enterprises with strict data-residency and sovereignty requirements that previously kept them from adopting third-party AI models at all.