ZEDEDA and SecEdge Partner to Strengthen Edge AI Security with Hardware-Rooted Protection

ZEDEDA, the leader in edge intelligence, and SecEdge, a provider of digital security solutions for physical AI infrastructure, announced a technology partnership to bring hardware-rooted security to ARM edge devices, closing one of the most significant gaps in enterprise edge AI deployments. The integration embeds SecEdge’s SEC-TPM firmware (fTPM) into ZEDEDA Edge Intelligence Platform, delivering field-updatable security that evolves with the threat landscape and activates on existing hardware, requiring no chip replacement or manufacturing disruption.

Gartner reports 76% of enterprises cite data privacy and security as their top AI risk. Edge environments introduce security challenges that go beyond traditional IT: devices operate in physically exposed locations, often without network perimeters, and are vulnerable to physical tampering, firmware replacement, and unauthorised access. The exposure is compounded by a fundamental hardware gap: most ARM-based devices lack a hardware TPM entirely, leaving organisations unable to verify device integrity, enforce secure boot, or meet emerging compliance requirements such as the EU Cyber Resilience Act.

Unlike a traditional hardware TPM chip, which is fixed at the time of manufacture and cannot be updated, a firmware TPM evolves alongside emerging threats. That includes support for next-generation encryption standards built to provide cryptographic agility and meet evolving security standards, something a physical chip cannot be updated to address. Because security functions run inside the main processor rather than on a separate component, the attack surface shrinks: there is no external chip to intercept, reset, or spoof. And because it deploys as firmware, it works on devices already in the field, not just new ones — eliminating hardware replacement costs and supply chain disruptions. As the NSA and CISA have identified, this kind of cryptographic device identity is foundational to supply chain integrity.

“Security at the edge has to start at the hardware level. If you can’t trust the device, you can’t trust anything running on it, including the AI models that are increasingly making autonomous decisions in production environments,” said Said Ouissal, CEO and founder of ZEDEDA. “ZEDEDA has always practised security by design, and our partnership with SecEdge extends that same hardware-rooted trust to ARM devices through firmware that evolves with the threat landscape, without requiring customers to replace hardware or change their manufacturing process. As autonomous edge agents become more prevalent, this kind of adaptive, future-proof trust layer is more important than ever.”

SecEdge’s SEC-TPM is a TCG TPM 2.0-compliant firmware TPM that provides hardware-anchored security without requiring a discrete TPM chip or a secure manufacturing facility. Integrated into the ZEDEDA Edge Intelligence Platform, it enables TPM-level device attestation, secure boot, and encrypted storage on ARM devices. ZEDEDA customers gain the same security posture across a broader range of edge hardware, with stronger compliance coverage and lower bill-of-materials costs.

“AI at the edge demands a new security foundation. As high-value AI models move to the edge and adversaries increasingly use AI to launch more sophisticated attacks, traditional security approaches are no longer sufficient,” said Sami Nassar, CEO of SecEdge. “SecEdge’s SEC-TPM, embedded within ZEDEDA’s platform, provides hardware-rooted security across the distributed device fleets enterprises are deploying today, establishing the trusted foundation required to securely scale edge AI.”

The partnership will expand beyond device security to include AI model protection at the edge, ensuring that AI models deployed on distributed devices remain secure, unaltered, and protected against unauthorised modification and cloning. This gives customers a consistent security posture across all of the edge devices in their fleet, ensuring that both the devices running AI workloads and the models deployed upon them remain secure and unaltered, without replacing hardware or adding to bill-of-materials costs.