Duality Technologies has announced a major step forward in privacy-focused artificial intelligence. The company now supports Google Cloud’s Confidential Computing platform, including NVIDIA GPU-powered confidential virtual machines. This allows large AI tasks, like training and running large language models, to be done securely on NVIDIA GPUs with full data protection. It’s a big move for companies that need to keep sensitive data private while using powerful AI tools.

Enhanced Performance for Secure AI Workloads

The update means that Duality’s platform can now run GPU-backed AI inference and encrypted retrieval-augmented generation within trusted environments. This results in much faster processing times compared to traditional CPU-only setups. Customers can combine complete data confidentiality with the high performance of NVIDIA’s H100 GPUs, opening up new possibilities for confidential AI applications that were once too slow or resource-intensive to consider.

With this support, organizations can now process sensitive workloads—such as medical data analysis or defense-related AI—without risking data leaks. The integration leverages Google Cloud’s Confidential Space and NVIDIA’s powerful GPUs, providing a seamless, secure environment for large-scale AI operations. The platform also supports Intel TDX and Google Cloud’s Key Management Service, making it versatile and easy to incorporate into existing workflows.

Impacts Across Industries

This breakthrough is especially important for regulated industries like healthcare, finance, and defense. These sectors often handle highly sensitive information and face strict compliance rules. Now, they can run complex AI models on confidential hardware without compromising security or performance.

Dr. Alon Kaufman, CEO of Duality Technologies, highlights the significance of this launch, saying, “This changes the game. Our customers can now run privacy-preserving AI with large language models at production scale.” Previously, secure AI workloads were limited to CPU environments, which were too slow for many real-world applications. The addition of GPU support makes high-throughput, privacy-preserving AI a practical reality.

This development paves the way for more advanced AI use cases, from real-time medical diagnostics to secure financial modeling. It also sets a new standard for how confidential computing can support the demanding needs of modern AI workloads. As Nelly Porter from Google Cloud notes, organizations can now process sensitive AI data entirely within trusted environments without sacrificing speed or efficiency. It’s a major milestone for the future of secure AI technology.

Inspired by

• https://ai-techpark.com/duality-technologies-enables-secure-genai-workflows-on-nvidia-gpus/