Curious about today's AI digest?ai-tldr.dev

Daily Digest

Pomegra Startups

Anthropic Licenses Opaque's Confidential AI Tech in 2026

Anthropic licensed confidential-AI technology from Opaque Systems and hired some of its engineers. Opaque continues as an independent company.

AIAcquisitionsNOTABLE4 min read
Anthropic Licenses Opaque's Confidential AI Tech in 2026

Anthropic has licensed Opaque Systems' confidential-AI technology and hired some of its engineers, including the CEO and CTO, while Opaque stays independent and its Series B valuation was about $300M.

Key Takeaways

  • Anthropic licensed Opaque's confidential-AI technology and hired some of its engineers, including CEO Aaron Fulkerson and CTO Rishabh Poddar.
  • Financial terms and the number of hires are undisclosed; Opaque raised a $24M Series B at about $300M in February.
  • Opaque continues as an independent company, so this is an acquihire-style deal, not an acquisition.

Lead

Anthropic agreed in mid-October 2026 to license technology from Opaque Systems, a five-year-old startup that secures sensitive data for use with AI agents, and to hire part of its team. The hires include CEO Aaron Fulkerson and CTO Rishabh Poddar. Neither company has disclosed the price of the license or how many engineers are moving. Opaque is not being bought, and it continues as an independent company.

What Did Anthropic Actually Get?

Anthropic received a license to Opaque's platform for running AI workloads on confidential data, plus people who know how to build it. The platform runs computation inside hardware-isolated environments, so the cloud operator cannot see the data being processed. It also produces cryptographic attestation that approved code ran on the data and followed set policies.

The technology began as a UC Berkeley RISELab research project about ten years ago. Opaque built it into a commercial product and worked with Nvidia, Intel, Microsoft and Google on confidential computing. It also gained customers in Europe and the United Arab Emirates.

Anthropic was already an Opaque customer before the deal. That history matters: the buyer had run the product and could judge what it was licensing.

Why Is This an Acquihire and Not an Acquisition?

The structure splits the technology from the company: Anthropic takes a license and a group of engineers, and the corporate entity stays outside its balance sheet. These deals have become a common way for large AI labs to absorb specialist teams without a full purchase. Anthropic gets the senior engineering talent, and Opaque keeps its remaining staff, contracts and cap table.

That distinction changes how to read the deal. A strategic acquisition would suggest Anthropic wanted Opaque's customer base. An acquihire suggests it wanted the people and the intellectual property, and it is not paying for the revenue.

What Does It Mean for Opaque's Customers?

Customers now face open questions about continuity. Opaque's listed customers include ServiceNow, Accenture and Encore Capital, and neither company has said whether the platform will keep serving them under the license-and-hire arrangement. Losing the CEO and CTO to a licensee is a heavy loss for a company of its size, and the remaining team will have to show it can keep shipping.

Fulkerson framed the agreement in a public statement as one that "carries OPAQUE's work forward." He said the company built a business from university research, helped establish confidential AI as a category and showed that some of the hardest security problems in AI could be solved.

How Does the February Round Look in Hindsight?

Opaque raised a $24M Series B in February 2026, led by Walden Catalyst, at a post-money valuation of about $300M. Intel Capital, Race Capital, Storm Ventures and Thomvest returned as investors, and the Advanced Technology Research Council joined as a new backer. Total funding reached $55.5M.

Eight months later, the company's founders are moving to a licensee. A $300M valuation on a $24M raise implies investors expected a standalone growth story. The license-and-hire outcome is a different path, and the price Anthropic paid will determine whether Series B holders see it as a good exit or a rescue. Because the terms are undisclosed, that cannot be judged yet.

Strategic Context

The deal fits Anthropic's push toward enterprises that want on-premises deployments, open models and sovereign AI architectures. Regulated buyers in finance, healthcare and government often will not send sensitive data to a shared cloud without proof of how it is handled. Attestation is the kind of evidence those buyers ask for, and building it in-house takes years of low-level hardware and cryptography work.

Competition in confidential computing is also tightening. Cloud providers and chipmakers already ship the underlying hardware features, which leaves the software layer that proves compliance as the contested part. Anthropic now holds a team that has built that layer.

Outlook

The deal gives Anthropic confidential-AI technology and senior engineers without the cost of a full acquisition, and it leaves Opaque as an independent company with a thinner leadership bench. Open items are the license price, the number of hires, and whether Opaque keeps serving ServiceNow, Accenture and its other customers. Expect Anthropic's enterprise product announcements to show how quickly the licensed technology reaches regulated buyers.

More Startup News