Quantum-resistant hardware security for Nvidia robotics platform

The result is a certified, quantum-resistant basis of trust, the so-called Root of Trust, on which future physical AI systems can be built. As robots and autonomous machines increasingly leave shielded industrial environments and operate in factories, logistics centers and public spaces, not only are security requirements increasing, but so are the economic risks. A cyber attack can result in business interruptions and liability claims that go far beyond classic data loss. For manufacturers and operators of robot systems, the choice of security architecture is therefore not a purely technical decision. It influences long-term competitiveness, the ability to obtain approval in regulated markets and the total cost of ownership over the entire product life cycle.

"Robots that perceive, analyze and respond to the real world are only as trustworthy as the security foundation on which they are built," said Dr. Stephan Zizala, Division President Connected Secure Systems at Infineon. "Infineon's Optiga TPM anchors a hardware-based trust foundation in the Nvidia Jetson Thor platform, which has already been proven in hundreds of millions of devices worldwide. The integration thus fulfills the special requirements of industrial robotics: long life cycles, real-time capability and reliable operation on a large scale. The integrated post-quantum cryptography ensures that this foundation is not only armed against current threats, but also offers protection over the entire lifetime of each robot."

"Physical AI systems operate in the real world, where security is a fundamental requirement," said Deepu Talla, Vice President Robotics and Edge AI at Nvidia. "Infineon's certified Optiga TPM for Nvidia Jetson Thor helps developers protect cryptographic keys, verify software integrity and deploy robot fleets securely at scale. This creates a hardware-based foundation of trust - the basis for secure and resilient autonomous systems."

The EU Cyber Resilience Act, the EU AI Act, the IEC 62443 standard for industrial systems and industry-specific standards in the healthcare and automotive industries are leading to new requirements for verifiable and auditable security at hardware level. This creates a demand driven by regulatory and compliance requirements that Infineon and Nvidia can address in a targeted manner.

Optiga TPM technology provides a physically isolated, FIPS and Common Criteria certified solution that is separate from the application processor. It enables measured boot and remote attestation, allowing operators and regulators to cryptographically verify that the software stack is authentic and unaltered at any point in a system's operation. It also provides hardware-protected storage for proprietary AI model keys, encrypted communication and cryptographically signed over-the-air updates.

The Optiga TPM, which Infineon claims is the industry's first TPM with a post-quantum secured firmware update mechanism, is designed as a root of trust that cannot be compromised even in the face of evolving cryptographic threats. Developers building Physical AI applications on Nvidia's Jetson Thor platform can rely on the hardware-based security foundation already anchored in the architecture phase and remain protected against current and future cryptographic threats in robotic systems.

The roadmap to complete post-quantum security is rounded off by Infineon's next-generation Optiga TPM. The TPM integrates algorithms such as ML-KEM and ML-DSA, which were standardized by the US National Institute of Standards and Technology (NIST) in 2024. Companies that build on the current Optiga TPM today will be able to make a smooth transition in the future. For the robotics industry, this is important beyond technical readiness. The regulatory framework for Physical AI is already moving towards mandatory PQC compliance. The architectural decision made at the beginning therefore determines whether a deployed robot fleet can meet these requirements over its entire service life or will be confronted with costly hardware interventions when the corresponding regulations come into force.