Duality Technologies

On September 23, we’re bringing together senior leaders from IBM, Red Hat, Intel, and Duality for a live executive discussion on what secure, sovereign AI looks like in practice. Featuring Alon Kaufman, Duality; Paul O'Neill O’Neill, Intel; Darren Orzechowski, Red Hat; Sachin Prasad, IBM, moderated by Ronen Cohen, Duality. Live attendees will also receive our Secure AI Across Boundaries Guide. Register to join us live. https://lnkd.in/dHmqApEY

Duality is heading to AFCEA International TechNet Europe 2026 in Berlin, September 15–16. We’ll be exhibiting at the event, and Mark T. will also be speaking. If you’re attending, we’d love to connect. Schedule some time with our team in advance, or stop by our booth while you’re there. Schedule a meeting → marketingteam@dualitytech.com

We’ve just published the 2026 Duality Guide to Privacy-Enhancing Technologies, designed to make the landscape easier to understand. The guide explains when each technology is useful, where the tradeoffs are, and how they can work together. That matters because PETs are moving beyond research and narrow use cases. They’re increasingly being used to support secure AI, encrypted search and analytics, federated AI, data sovereignty, and collaboration across organizations and jurisdictions. If you’re trying to figure out how to use sensitive or regulated data for AI and analytics without unnecessarily exposing it, this is a good place to start. Download the 2026 PETs Guide: https://lnkd.in/d4MKQYxD #PrivacyEnhancingTechnologies #PETs #SecureAI #DataPrivacy #FederatedAI #ConfidentialComputing

What does it take to make fully homomorphic encryption fast enough for real-world workloads? A big part of the answer comes down to something very familiar to hardware and performance engineers: large-integer arithmetic. In the next OpenFHE webinar, researchers from NVIDIA, Fair Math, Inc, and Duality Technologies will look at the math and implementation challenges behind vectorized FHE schemes and how GPU-accelerated libraries such as NVIDIA cuPQC could help push performance forward. This session is especially relevant for engineers and researchers working in high-performance computing, GPU acceleration, cryptography, numerical computing, and systems optimization who are curious about applying that expertise to FHE. Featuring Yarkin Doroz, NVIDIA; Gurgen Arakelov, Fair Math; and Yuriy Polyakov, Duality Technologies. This is the third webinar in our series exploring major projects in the OpenFHE ecosystem. https://lnkd.in/eqUsDxHu #OpenFHE #FHE #GPUComputing #HPC #Cryptography #NVIDIA

On September 23, we’re bringing together senior leaders from IBM, Red Hat, Intel, and Duality for a live executive discussion on what secure, sovereign AI looks like in practice. We’ll explore: - How to collaborate across organizations and jurisdictions without giving up control of sensitive data - How to use cloud and GPU resources for sensitive AI workloads - How to incorporate external data while protecting proprietary models, logic, and IP - How governance, infrastructure, confidential computing, and privacy-enhancing technologies fit together Featuring Alon Kaufman, Duality; Paul O'Neill O’Neill, Intel; Darren Orzechowski, Red Hat; Sachin Prasad, IBM, moderated by Ronen Cohen, Duality. Live attendees will also receive our Secure AI Across Boundaries Guide. Register to join us live. https://lnkd.in/dHmqApEY #SecureAI #SovereignAI #ArtificialIntelligence #DataPrivacy #ConfidentialComputing

Exciting news: IBM Sovereign Core and Duality are working together to help organizations securely collaborate across sensitive, distributed data. By bringing together IBM’s sovereign infrastructure with Duality’s privacy-enhancing technologies, organizations can run AI and analytics across organizational, jurisdictional and cloud boundaries while maintaining governance, security and operational control. For regulated industries, this creates powerful new possibilities for how sensitive data can be used across departments, partners, clouds and jurisdictions without compromising the controls that sovereignty requires. We’re proud to work with IBM on a shared vision for the next era of enterprise AI, and excited to see where this collaboration leads. Thank you to Sachin Prasad, Will Streit and the IBM Sovereign Core team for the support on this. Read the full article: https://lnkd.in/evSTkCA5 #SovereignAI #EnterpriseAI #SecureAI #PrivacyEnhancingTechnologies

The White House Office of Science and Technology Policy released an exciting revision of the National Security Science & Technology Strategy this week (https://lnkd.in/gr-3w6kb). This is the first document to satisfy the CHIPS and Science Act requirement for a national security S&T strategy. There is a lot to unpack in this document, and I'll be giving a really close read, especially in the context of the recently preceding Presidential budget request for FY2027 (https://lnkd.in/gVZX52_e). I'm already seeing commentary focused on the (really important) four pillars in the document, but some especially consequential material is in Appendix A, the updated Critical and Emerging Technologies (CET) list. Several entries are worth noting that I think are going to be increasingly important for operational and commercial impact, but are especially telling in that they were identified so prominently: -> "Distributed and privacy-preserving machine learning" is now a named subfield under AI and autonomy. To my knowledge, this is the first time Privacy-Preserving Machine Learning (PPML) appears by name in a strategy document at this level. -> Post-Quantum Cryptography (PQC) is listed under information management and cybersecurity. ->"Enabling technologies for high security computing" appears under future computing. ->Low-SWAP-C embedded computing for the tactical edge is also highlighted. It's a great time to be focused on building capabilities, products and companies around these technologies! The significance of the CET designations shape where federal R&D funding, program offices, and procurement attention concentrate over the following several years. For the secure and embedded computation community, this is the document to read *very* closely. I expect much more to come from the leading sponsors and investors such as Defense Innovation Unit (DIU), SOFWERX, IQT, Office of Strategic Capital, and much more. I'm excited and I'll be publishing a series of longer analyses on my Substack examining what the strategy asks of technology and markets. #PQC #PPML #DefenseTech https://lnkd.in/gr-3w6kb

Google just made an announcement that has been part of the long march of encrypted computing maturation and adoption. Their new Google HEIR compiler converts pre-trained AI models to run on encrypted data, greatly reducing the technical effort need, with demos built alongside hardware accelerator companies like Niobium, Belfortt, Cornami, Inc., and Optalysys. Google's framing: homomorphic encryption is "rapidly maturing," and the trade-off is now cost, not capability. And the cost is falling fast. When we started this work under DARPA programs that invested in Fully Homomorphic Encryption (#FHE) in the early 2010s, FHE was originally a mathematical curiosity, millions of times too slow. Over a decade, a layered ecosystem emerged: open-source libraries (including OpenFHE, which I co-founded and which compilers like HEIR target as a backend), then hardware acceleration (Defense Advanced Research Projects Agency (DARPA)'s DPRIVE program seeded much of today's accelerator landscape). IARPA made some fundamental early investments in compilers for FHE, but the field has moved forward and now compilers like HEIR let a much wider range of developers use it all without a cryptography team. Duality Technologies has been deploying this operational FHE-protected secure data collaboration software stacks for governments and regulated enterprises for years. We're happy to support and be a part of the open-source ecosystem for FHE, along with Google and the many other collaborators and contributors in the domain. Alon Kaufman Rina S. shafi goldwasser Vinod Vaikuntanathan Yuriy Polyakov Andreea Alexandru Ahmad Al Badawi, PhD https://lnkd.in/ggz8CtDh

The research sparked some great questions. Now we’re bringing the teams together to dig into them and inviting the broader community to join the conversation. Researchers from NVIDIA, Fair Math, Inc, and Duality Technologies recently published “Efficient Large-Integer Arithmetic for FHE,” exploring the algorithms and implementation techniques behind one of the key building blocks of high-performance FHE. On our next OpenFHE webinar, researchers from all three teams will come together to discuss the work, where it goes next, and how others can get involved. Yarkin Doroz (NVIDIA), Gurgen Arakelov (Fair Math), and Yuriy Polyakov (Duality Technologies) will discuss: • The key findings from the joint research • How GPU-accelerated cryptographic math libraries such as NVIDIA cuPQC can accelerate FHE • The work already underway to build more efficient large-integer kernels for OpenFHE • The research and engineering challenges that still need to be solved—and opportunities for the community to contribute This is the third webinar in our series highlighting major projects around OpenFHE, and this one is particularly relevant for engineers and scientists working in high-performance computing, hardware acceleration, and cryptographic mathematics even if FHE isn’t yet your area of expertise. 📄 Read the paper: https://lnkd.in/eC7xxMvp 🔗 Register for the webinar: https://lnkd.in/eqUsDxHu

The interesting thing about the Bridging the Privacy & Security Gap category is that it gets at a problem many organizations still haven’t fully solved. Security teams are focused on protecting sensitive data. Privacy teams are focused on making sure it is used appropriately. But increasingly, the real challenge is doing both while still making that data useful. That’s exactly the space Duality was built for. We’re very pleased to be named a finalist in this category at the 2026 PICCASO (Privacy, InfoSec, Compliance, Culture, Ai, Strategy & Ops) Awards. A meaningful recognition for a problem we think is only becoming more important. #PICCASOAwards #PrivacyEnhancingTechnologies #DataSecurity #DataPrivacy #SecureAI