OPAQUE, the Confidential AI company headquartered in San Francisco, California, announced it has acquired advanced cryptographic AI technologies from the Technology Innovation Institute (TII), the applied research pillar of Abu Dhabi’s Advanced Technology Research Council (ATRC). The acquired technology – already proven in real-world use cases – adds two critical capabilities to OPAQUE’s platform: confidential AI model training powered by advanced cryptographic techniques such as multi-party computation and fully homomorphic encryption, as well as post-quantum cryptographic protections.
Founded by researchers from UC Berkeley’s RISELab, and now with this acquisition, OPAQUE supports Confidential AI workflows across training and inference, moving enterprises from isolated AI experiments to production deployment 4-5x faster.
The acquisition was overseen by H.E. Faisal Al Bannai, Adviser to UAE President His Highness Sheikh Mohamed bin Zayed Al Nahyan and Secretary General of ATRC, and Ion Stoica, Co-Founder and Board Member of OPAQUE, Co-Founder of Databricks, and Professor of Computer Science at UC Berkeley. It marks the first time cryptographic AI technologies developed in the UAE have been acquired and deployed at scale by a US-based technology company.
Also Read: Sapio Sciences Brings Claude Cowork to the Lab
“This acquisition reflects what the UAE has set out to do – produce foundational technology that the world adopts,” said H.E. Faisal Al Bannai, Adviser to the UAE President and Secretary General of ATRC. “Cryptographic AI capabilities developed in Abu Dhabi are now being deployed at global scale by a leading US technology company. The UAE is not only adopting the AI economy – we are leading. ATRC will continue to invest in the research that defines the next generation of trusted AI.”
“The future of AI depends on unlocking the data organizations have never been able to touch,” said Ion Stoica, Co-Founder of OPAQUE. “Most enterprises sit on a corpus of data too sensitive to use and too valuable to ignore. With this acquisition, OPAQUE is the only platform delivering hardware-attested cryptographic evidence across the full AI lifecycle – training, fine-tuning, inference, and agents – with protections engineered to withstand quantum-era threats. That combination doesn’t exist anywhere else in the market today.”
“AI agents are extraordinarily powerful: they operate at machine speed with human-like capabilities, delivering in days what entire teams would struggle to complete in a year,” said Aaron Fulkerson, Chief Executive Officer of OPAQUE. “That same power is exactly why they are so difficult to move into production. An agent connected to sensitive systems or regulated data — patient records, clinical research, financial claims – can do damage in days that a team of malicious actors couldn’t match in a year. The only way to safely deploy them is with hardware-enforced, verifiable rules: provable evidence of what ran, where it ran, and which rules were enforced. Building on primitives from every major cloud provider, OPAQUE makes it possible to move agents into production – even on the most sensitive and most regulated systems and data.”
Enterprises sit on vast amounts of sensitive data — patient records, financial transactions, classified intelligence, proprietary research – that could transform their AI capabilities. But using that data across the full AI lifecycle has required stitching together point solutions from multiple vendors, each covering a different phase: one for training, another for inference, another for agent workflows. The result is gaps, complexity, and risk that compliance teams won’t sign off on.
OPAQUE eliminates those gaps. The acquired technology extends OPAQUE’s confidential AI platform across training, fine-tuning, inference, and AI agent execution. It delivers verifiable guarantees at every stage, backed by post-quantum cryptography that protects workloads against current and future threats. For example, ServiceNow runs OPAQUE in production to extend AI capabilities to its enterprise customers without exposing their data. A national healthcare system can train diagnostic models on patient data across jurisdictions, run inference across facilities, and deploy AI agents on live clinical data – all on a single platform. The platform generates hardware-attested evidence aligned with SOC 2, ISO 27001, ISO 42001, GDPR Article 32, and EU AI Act high-risk obligations, proving data remained private, policies were enforced, and regulatory requirements were met.
SOURCE: TII