Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsAnthropic has not been confirmed to have bought OPAQUE outright. A deal reported on October 13, 2026 combines a license to OPAQUE’s confidential-computing technology with the hiring of part of its team, including CEO Aaron Fulkerson and CTO Rishabh Poddar. Financial terms were not disclosed.
What Anthropic and OPAQUE agreed to
Constellation Research reported on October 13, 2026 that Anthropic entered a new agreement with OPAQUE, a company focused on confidential computing and confidential AI. The reported structure has two parts: Anthropic is licensing OPAQUE technology, and some OPAQUE employees are joining Anthropic. The named hires include OPAQUE chief executive Aaron Fulkerson and chief technology officer Rishabh Poddar.
As an Amazon Associate I earn from qualifying purchases.
That makes this an acquihire-style licensing and hiring transaction, not a confirmed full-company acquisition. Neither the purchase price nor other financial terms were disclosed. There is also no public statement establishing that OPAQUE as a corporate entity has been absorbed into Anthropic.
| Question | What is established |
|---|---|
| Did Anthropic buy OPAQUE? | No outright acquisition has been confirmed in the available announcement. |
| What is Anthropic obtaining? | A license to OPAQUE’s confidential-AI technology and selected team members. |
| Which leaders were named? | OPAQUE CEO Aaron Fulkerson and CTO Rishabh Poddar. |
| How much did the deal cost? | Not disclosed. |
What “confidential AI” means
Confidential AI is designed to let an organization process sensitive data while reducing the ability of cloud operators, infrastructure administrators or unauthorized software to inspect that data or the resulting model activity. It combines protected execution environments with cryptographic controls and verifiable policy enforcement.
#1 Best Overall
Hardware attestation
Hardware attestation allows a workload to prove that it is running inside an approved, protected environment. An enterprise can check the identity and configuration of the computing environment before releasing data or model keys. The goal is to make security policy verifiable rather than relying only on a provider’s written assurances or software configuration.
Cryptographic protections
OPAQUE says its platform uses several advanced techniques:
- Multi-party computation: multiple parties can jointly compute on inputs without any one participant receiving the complete underlying data.
- Fully homomorphic encryption: calculations can be performed on encrypted data, allowing results to be produced without exposing the plaintext to the processing system.
- Post-quantum cryptography: encryption and key-management approaches intended to remain resistant to attacks from future quantum computers.
These methods have different performance, cost and implementation trade-offs. Their presence in a platform does not automatically make every deployment private or compliant; the protections still depend on correct configuration, key management, identity controls and the organization’s operating procedures.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Rank #2
OPAQUE’s stated AI lifecycle coverage
OPAQUE describes its technology as covering the full AI lifecycle rather than only a single inference endpoint.
| Lifecycle stage | What confidential controls are intended to protect |
|---|---|
| Training | Sensitive datasets and the infrastructure processing them. |
| Fine-tuning | Proprietary examples, domain knowledge and resulting model updates. |
| Inference | User prompts, retrieved records and model outputs during production use. |
| AI-agent execution | Tools, credentials, workflows and private information an agent can access while acting. |
Covering agents is particularly important because an agent may call databases, internal software and external services, creating more opportunities for accidental disclosure than a single question-and-answer request. A lifecycle approach also gives security teams one architectural model for experimentation and production instead of treating each stage as a separate exception.
Why regulated organizations are interested
Healthcare providers, banks, insurers, government agencies and defense contractors often hold data that cannot be sent to an ordinary shared service without strict contractual, legal and technical controls. Their barriers are not limited to model accuracy. They must also address leakage of personal or classified information, audit requirements, data-residency rules, customer confidentiality and restrictions on who may administer the infrastructure.
Rank #3
Confidential computing can help address those concerns by limiting plaintext exposure and producing evidence about the environment in which a workload ran. It can support architectures in which sensitive processing stays in a customer-controlled facility, a designated national region or an approved sovereign cloud. Those properties are relevant to Anthropic’s enterprise push, including the on-premises and sovereign-AI architectures that Constellation Research identified as areas of interest.
Free tools Windows power users keep installed
One-click scans. No signup required.
Confidential controls do not remove every risk. An authorized application can still misuse data, a user can still copy an output, and poor access policies can still grant an agent excessive privileges. Organizations must continue to classify data, restrict tools, monitor activity and define retention rules.
How the move fits Anthropic’s enterprise strategy
Frontier models are easier to adopt when a company can send ordinary business information to a hosted API. Adoption slows when the same company wants to use patient records, trading information, government files or defense data. Licensing OPAQUE’s technology and adding specialists who built it gives Anthropic a way to strengthen the infrastructure layer around its models, not just the models themselves.
The arrangement could support deployments where a customer needs proof that data and model operations stayed inside an approved boundary. It could also make confidential controls available across training, fine-tuning, inference and agents, rather than offering protection only for one API call. The public information does not establish a product launch date, supported hardware list, customer roster or specific Anthropic service that will incorporate OPAQUE technology.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.OPAQUE’s funding and performance claims
OPAQUE reported a $24 million Series B financing at a $300 million valuation in 2026. Those figures describe OPAQUE’s funding context before the new Anthropic agreement; they do not reveal what Anthropic paid, if anything, for the license or hiring arrangement.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →OPAQUE also said its platform can move enterprises from isolated AI experiments to production deployment four to five times faster. That is a company claim, not an independently verified benchmark. No test design, baseline, workload mix or third-party validation was disclosed alongside the figure, so it should not be treated as a guaranteed implementation result.
What the executives said
“We’re proud to share a new agreement with Anthropic that carries OPAQUE’s work forward. We’re excited to see its impact continue in shaping how AI is built and secured,” Fulkerson said.
OPAQUE co-founder Ion Stoica framed the broader objective this way: “The future of AI depends on unlocking the data organizations have never been able to touch.”
What to watch next
- Whether Anthropic identifies the products, model services or deployment environments that will use OPAQUE technology.
- Whether customers can run the resulting capabilities on premises or in sovereign regions, and which jurisdictions and hardware are supported.
- Independent evidence about performance, cost and operational overhead when cryptographic protections are enabled.
- How Anthropic handles key ownership, administrator access, logging, data retention and agent permissions.
- Whether OPAQUE continues as a standalone business, operates under a new Anthropic unit or becomes a technology layer inside existing services.
For now, the clearest reading is narrower than an acquisition headline: Anthropic is adding confidential-computing technology and experienced personnel to help make sensitive enterprise AI workloads more deployable, while the ownership structure and commercial terms remain private.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
