By 2027, your factory’s autonomous forklift will need a quantum-resistant cryptographic passport to get liability insurance.
A scribe verifies an immutable ledger with a seal, symbolizing remote attestation and measured boot in quantum-secure robotics.
On June 3, 2026, Infineon Technologies announced the integration of its OPTIGA TPM SLB 9672 security chip with NVIDIA’s Jetson Thor platform. The move provides a hardware-based, quantum-resilient root of trust for autonomous mobile robots, humanoids, and physical AI systems. The announcement comes ten months after Infineon and NVIDIA revealed a separate collaboration to enable humanoid robots with precise motion and efficiency, combining Infineon microcontrollers, sensors, and actuators with Jetson Thor modules. This new integration adds a cryptographic anchor to that hardware stack. It also arrives as the European Union’s Cyber Resilience Act enforcement deadlines draw closer, making auditable hardware security a procurement requirement rather than a technical nice-to-have.
Deepu Talla, NVIDIA’s vice president of robotics and edge AI, framed the integration as a foundational layer. “Physical AI systems operate in the real world, where security is foundational,” Talla stated. “Infineon’s certified OPTIGA TPM for NVIDIA Jetson Thor helps developers protect keys, verify software integrity and securely provision robot fleets at scale.”
An underwriter examines a document's seal of authenticity, representing the coming insurance mandate for hardware security in robotics.
The Threat and the Regulatory Hammer
Autonomous systems deployed in factories, warehouses, and hospitals carry operational lifecycles stretching beyond a decade. That longevity creates a “harvest now, decrypt later” attack surface: an adversary can intercept and store encrypted telemetry, firmware updates, or model weights today, then decrypt that data once cryptographically relevant quantum computers arrive. The NIST post-quantum cryptography standardization effort, which produced the ML-KEM and ML-DSA algorithms, addresses this timeline mismatch. Infineon’s TPM roadmap includes support for both.
What forces the issue is not the distant quantum threat. The EU Cyber Resilience Act, the EU AI Act, and IEC 62443 for industrial control systems now demand demonstrable, auditable hardware-level security. A software-only approach leaves no tamper-resistant root of trust. For a robotic arm operating next to human workers or an autonomous forklift navigating a shared warehouse floor, a compromised software stack translates directly to kinetic damage. Insurers and risk auditors are beginning to notice.
Infineon is not treating this as a single-product play. The company is embedding post-quantum cryptography across its microcontroller portfolio. Its PSOC Control C3 Performance Line MCUs comply with the NSA’s Commercial National Security Algorithm Suite 2.0 for post-quantum firmware protection. The AURIX and TRAVEO families, used throughout automotive supply chains, are receiving the same treatment. The Jetson Thor integration is the most visible node in a broader push to make PQC-ready silicon a default component.
How the TPM-Jetson Thor Mechanism Works
The OPTIGA TPM SLB 9672 functions as an autonomous cryptographic vault. It is certified under FIPS 140-3 and Common Criteria guidelines. Its integration with Jetson Thor enables five specific capabilities: secure key storage, measured boot, remote attestation, encrypted communications, and protected software updates.
Here is the operational sequence that matters for compliance auditors. During startup, the platform generates cryptographic hashes of the operating system and AI model weights. A remote monitoring console can query those hashes and verify that the runtime software stack remains unmodified. If an attacker replaces a perception model with a poisoned version, the hash mismatch triggers an alert before the robot moves. The module also uses a post-quantum cryptography secured firmware update mechanism to protect the device’s root authority against future quantum decryption vectors.
Dr. Stephan Zizala, Division President of Connected Secure Systems at Infineon, summarized the architecture: “Robots that sense, think and act in the real world are only as trustworthy as the security foundation they are built on.” The TPM provides that foundation as a physical chip. It cannot be patched over by a compromised hypervisor or container runtime.
The Frontier Take
The consensus reads the Infineon-NVIDIA announcement as a post-quantum cryptography story. It is actually a regulatory-compliance infrastructure story where “quantum” is the marketing hook. The NIST PQC algorithms are on a roadmap. They are not yet fully deployed in the TPM firmware. What ships now solves today’s classical intrusion and ransomware problems: a FIPS-certified, measured-boot, attested-update pipeline that provides immediate auditability. The EU Cyber Resilience Act’s enforcement deadlines are not distant. They are arriving. Infineon is selling cryptographic proof that a robot booted into a known-good state. That proof is what enterprise risk auditors and commercial insurers will demand.
Here is the prediction, stated plainly. Within 12 to 24 months, commercial insurers and enterprise risk auditors will begin explicitly requiring FIPS 140-3 and PQC-ready hardware roots of trust for any autonomous mobile robot or humanoid deployed in shared human spaces. The reasoning is straightforward. A fleet operator without cryptographic attestation cannot prove to an underwriter that its robots are running authorized software. Without that proof, liability coverage becomes unpriceable. OEMs that treat security as a software bolt-on will find themselves locked out of enterprise RFPs and liability coverage. The bifurcation will be brutal. Fleets with a hardware root of trust will be insurable. Fleets without one will be uninsurable.
This dynamic will trigger a wave of last-minute hardware redesigns. Robotics OEMs that shipped products without a TPM slot on the mainboard will scramble to retrofit or redesign. At least one major robotics bankruptcy or product recall will trace directly to an unsecured fleet exploit that causes injury or property damage. The Infineon-NVIDIA integration, combined with the August 2025 humanoid robotics collaboration, positions the two companies to capture the compliance-driven rebuild.
What This Means for the Industry
Robotics OEMs should audit their bill of materials now. If the mainboard lacks a discrete TPM or a secure element with equivalent certification, the design has a shelf life measured in months, not years. Fleet operators evaluating autonomous systems for deployment in 2027 and beyond should demand cryptographic attestation capabilities in their next RFP. A vendor that cannot demonstrate measured boot and remote attestation is selling a liability time bomb.
Investors screening robotics startups should treat a hardware root of trust as a moat indicator. A startup that embedded a certified TPM early will clear enterprise procurement hurdles faster than a competitor that must redesign its compute module. The TPM is not a differentiator. It is becoming a mandatory line item, like a seatbelt.
By 2027, that autonomous forklift will not move without a quantum-resistant cryptographic passport. The passport is a FIPS-certified TPM. The checkpoint is the insurance underwriter’s audit. The companies that install the checkpoint now will survive the bifurcation. The ones that wait will watch their fleets get parked.
