Anthropic acquires confidential-computing startup Opaque Systems
Anthropic has acquired Opaque Systems, a startup focused on confidential computing for AI, according to AI Weekly. The deal is described as a license-and-hire that brings Opaque’s CEO, CTO and some engineers to Anthropic, while Opaque says its operations are winding down.

Key takeaways · 4
- 01
AI Weekly describes the acquisition as a license-and-hire involving Opaque CEO Aaron Fulkerson, CTO Rishabh Poddar and some engineers—not all Opaque employees.
- 02
Financial terms were not disclosed in AI Weekly’s report.
- 03
OPAQUE says its operations are winding down, but five named projects remain open and available under their existing licenses.
- 04
OPAQUE says its technology protects AI workloads on sensitive data and its verifiable-AI work aims to provide cryptographic evidence of execution and policy.
What the deal includes
AI Weekly reports that Anthropic acquired Opaque Systems and describes the arrangement as a license-and-hire.[1] The report says Opaque CEO Aaron Fulkerson, CTO Rishabh Poddar and some engineers are joining Anthropic, but not all Opaque employees.[1] OPAQUE’s own announcement says its team is joining Anthropic to carry its work forward, and that Opaque Systems is winding down its operations.[2][3] The financial terms were not disclosed in AI Weekly’s account, so the available evidence does not establish a purchase price or other deal economics.[1]
The technology Opaque describes
OPAQUE says its technology lets organizations run AI workloads on sensitive data with hardware-enforced protections.[2] It describes its verifiable-AI work as aiming to provide cryptographic evidence of what ran, where it ran and under what policies.[2] OPAQUE says organizations needed a way to independently verify how AI systems and agents behaved.[2] The company also says its work included the Agent Governance Toolkit, TRACE, Confidential MCP and Weight Custody Manifest.[2] These are descriptions from OPAQUE; the evidence does not provide an independent assessment of the technology’s performance or adoption.
What remains available
OPAQUE says its open-source work continues at AgenTrust.[3] It specifically says the Weight Custody Manifest, Agent Manifest, Confidential MCP, Confidential A2A and TRACE remain open and available under their existing licenses.[3] That gives teams considering these projects a stated path to access them even as Opaque Systems winds down its operations.[3] The announcement does not specify how future maintenance or support will work, so organizations should check the projects’ current documentation and licensing before relying on them. The latter is a practical next step, not a commitment made in OPAQUE’s announcement.
Background and reported context
OPAQUE says the company began as a UC Berkeley RISELab research project ten years ago and brought confidential computing from research into production AI environments.[2] Race Capital also lists ServiceNow, Accenture and Anthropic among Opaque’s customers.[4] Separately, AI Weekly reports that Opaque raised a $24 million Series B at a $300 million valuation; those figures are not the disclosed terms of the Anthropic deal.[1]
For teams working with sensitive data, OPAQUE says its technology lets organizations run AI workloads on sensitive data with hardware-enforced protections, and its verifiable-AI work aims to provide cryptographic evidence about execution and policy. OPAQUE says its named open-source projects remain available under their existing licenses. The evidence confirms project availability but does not establish maintenance or support arrangements; AI Weekly reports that the deal’s financial terms were not disclosed.
Why it matters
Put this to work — one session a day, built for your industry.
Create a free account for a daily session — eight questions and one real-work challenge, on the news that affects your role.
Start freeHow this developed
20 October 2026
Anthropic acquires confidential-computing startup Opaque Systems