Anthropic Acquires Startup Opaque Systems

Anthropic, the developer of the Claude model, has acquired the startup Opaque Systems, which specializes in confidential computing for artificial intelligence. The deal apparently includes a licensing agreement for the technology and the hiring of key Opaque personnel, including CEO Aaron Fulkerson and CTO Rishabh Poddar. This acquisition could enhance Anthropic's capabilities in handling sensitive data and expand its presence in segments with strict privacy requirements.

Opaque Systems emerged from RISELab, a research lab at the University of California, Berkeley. The company focuses on confidential computing technologies that allow AI workloads to be run securely: data is processed in such a way that neither the infrastructure operator nor the cloud provider gains access to it. This area is particularly important for Anthropic, as its clients increasingly operate in regulated industries—finance, healthcare, government—where data leaks or misuse are unacceptable.

The nature of the deal is noteworthy in itself. The phrasing about technology licensing and the transition of key employees, rather than a classic acquisition with full business integration, is increasingly common in AI startup deals: major players gain access to developments and the team without burdening themselves with the entire corporate structure of the acquired company. In the case of Opaque, this means Anthropic not only gains the technological base of confidential computing but also the expertise of the people who created it.

In light of the deal, market valuations have shifted significantly. According to the provided data, the likelihood that Anthropic will have the best AI model by the end of October 2026 has increased to 28 percent. This reflects growing expectations regarding how the company will utilize the acquired assets and strengthen its position in the race for large models.