Anthropic Acquires $300M Opaque Systems to Verify AI Agents

Anthropic has acquired Opaque Systems, a startup specializing in confidential computing and verifiable AI technology, according to SecurityWeek. The deal allows Anthropic to integrate Opaque’s team and technology into its own operations, marking a strategic move to enhance security and verification capabilities for artificial intelligence systems. Opaque Systems announced the acquisition last week, stating that its employees are joining Anthropic. The company is winding down its independent operations as part of the transition.

How it unfolded

• Opaque Systems announced the acquisition by Anthropic last week, confirming its team is joining the AI developer.
• The company raised $24 million in a Series B funding round earlier this year, achieving a $300 million valuation.
• Opaque Systems began as a research project at UC Berkeley’s RISELab approximately ten years ago.

Who is affected

Opaque Systems listed Anthropic, ServiceNow, and Accenture among its customers prior to the acquisition. The startup had developed technology used in production AI deployments, focusing on protecting sensitive data and models through hardware-based security. With the acquisition, Anthropic gains access to Opaque’s expertise in confidential computing and its recent work on verifiable AI. This includes tools designed to provide cryptographic evidence of AI system behavior, addressing growing concerns about trust and security in automated agents.

What happens next

Opaque Systems’ website now states that the company is winding down its operations. However, the company has not specified what this transition means for its existing customers beyond Anthropic. The acquisition allows Anthropic to carry forward Opaque’s work, potentially integrating its verifiable AI technologies into future products. Financial details of the deal were not disclosed. The move signals a broader industry focus on securing AI systems through cryptographic verification rather than relying solely on traditional trust assumptions.

The background in plain language

Confidential computing uses hardware security to protect data while it is being processed, ensuring that even cloud providers cannot view sensitive information. Verifiable AI extends this by using cryptography to prove exactly what code ran and under what rules. Teams usually check whether vendors provide cryptographic evidence of system behavior. This approach replaces blind trust with verifiable proof, helping organizations ensure AI agents act within defined policies and security boundaries without revealing proprietary data.

What to do and how to stay safe: Anthropic

• Review current AI vendor contracts to determine if they provide cryptographic evidence of system behavior and data handling.
• Assess whether sensitive data processing relies on hardware-based security features that protect information during computation.
• Monitor vendor announcements regarding changes in data protection methodologies or shifts from confidential to verifiable computing models.
• Verify that AI systems operating on sensitive data have independent mechanisms to confirm they are running under approved security policies.

Step-by-step guide: Injection attacks examples: how prompt leaks break AI systems

General security guidance from the Circuit Observer newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

What did Anthropic acquire from Opaque Systems?

Anthropic acquired Opaque Systems, a startup that developed technology for confidential computing and verifiable AI, including its team and intellectual property.

How much was Opaque Systems valued at before the acquisition?

Opaque Systems was valued at $300 million after raising $24 million in a Series B funding round earlier this year.

Who are the founders of Opaque Systems?

The founders include Ion Stoica, co-founder of Databricks, and Raluca Ada Popa, an encryption and systems security researcher now at Google DeepMind.