Fortanix Inc., a data security provider, and NVIDIA have introduced a joint solution designed to offer a secure, on-premises platform for running sovereign, agentic AI in regulated environments and AI Factories.
Industries such as finance, healthcare, and government frequently handle sensitive data, including personally identifiable information, healthcare records, financial transactions, and classified material. These organizations often face challenges in leveraging AI due to the risks associated with moving such data to cloud environments. The collaboration aims to mitigate these risks by enabling secure AI innovation without compromising data sovereignty or compliance.
The joint platform integrates Fortanix’s confidential AI pipeline, composite attestation, and HSM-gated key release with NVIDIA Confidential Computing. This enables enterprises to deploy AI agents on sensitive data, leveraging NVIDIA Hopper and NVIDIA Blackwell GPU Architectures. The solution is built for sectors where confidentiality and compliance are critical, allowing organizations to adopt AI without sacrificing trust or performance. Customers can also integrate this solution into existing AI factories to enhance security and meet regulatory requirements.
Anuj Jaiswal, chief product officer at Fortanix, stated, “Enterprises in finance, healthcare and government want to harness the power of AI, but compromising on trust, compliance, or control creates insurmountable risk. We’re giving enterprises a sovereign, on-prem platform for AI agents—one that proves what’s running, protects what matters, and gets them to production faster.”
Justin Boitano, vice president of enterprise AI products at NVIDIA, added, “AI factories are engines of enterprise transformation, and must be built on a foundation of security and trust. Together with Fortanix, we’re bringing NVIDIA Confidential Computing into these next-generation environments, enabling organizations to run powerful AI with verified confidentiality, compliance and control in on-premises data centers.”
Key features of the Fortanix platform include end-to-end trust, ensured by protecting data and models throughout their lifecycle with NVIDIA Confidential Computing and composite attestation across CPUs and GPUs, leveraging NVIDIA NRAS. Cryptographic keys are released by Fortanix Data Security Manager (DSM) only after workloads are verified, establishing an “attestation-gated” approach for sensitive AI operations. Compliance is integrated by design, with the platform backed by a FIPS 140-2 Level 3 HSM, enforcing strict key custody, role-based access controls, and consistent audit logs to support compliance demonstration and reduce operational risk. The Fortanix Armet AI solution is designed to accelerate time-to-production by providing a turnkey on-premises AI platform with built-in AI guardrails, observability, and orchestration.
Fortanix will demonstrate the new platform at NVIDIA GTC, scheduled for October 27–29, 2025, at the Walter E. Washington Convention Center in Washington, D.
C. Attendees can visit Fortanix at booth I-7.
Fortanix is a leader in data-first cybersecurity and confidential computing, offering a unified platform that secures sensitive data across on-premises and multi-cloud environments. Its technology uses advanced encryption and key management, including quantum-resistant encryption, to secure data at rest, in transit, and in use. The company’s approach supports running applications and AI workloads within secure hardware enclaves, a method known as Confidential Computing, supported by technology firms including Intel, Microsoft, and NVIDIA.
