Can on-device runtime security become the standard for AI infrastructure in the age of edge computing?

Why is on-device runtime security emerging as a critical requirement for AI infrastructure deployed at the edge?

Exein SpA, the Italian embedded cybersecurity specialist, had recently secured a €70 million Series C funding round led by Balderton Capital, with participation from Supernova and Lakestar. The investment, which values Exein at approximately €500 million, is earmarked to expand its “digital immune system” technology beyond IoT devices into artificial intelligence infrastructure and large language models (LLMs) running on edge computing hardware. Exein projects €20 million in revenue for 2025, supported by a 450 percent year-over-year growth rate, signaling increasing institutional confidence in embedded security for edge-based AI systems.

Founded in 2018, Exein has rapidly scaled its runtime security technology to protect over one billion connected devices globally. As enterprises shift AI workloads closer to the edge for lower latency and improved privacy, the need for on-device, real-time threat detection is gaining urgency. Analysts are questioning whether embedded runtime security could soon become the default safeguard for AI infrastructure in high-risk and regulated industries.

What makes on-device runtime security distinct, and why is it particularly suited for AI workloads at the edge?

On-device runtime security integrates monitoring and response mechanisms directly into the firmware or operating environment of a device, functioning continuously during execution. Unlike traditional cloud-based or perimeter-focused models, runtime security detects behavioral anomalies at the point of execution, isolating threats in real time without external network dependence.

Exein has optimized this approach for constrained devices by building a lightweight, Rust-based runtime agent capable of syscall-level monitoring, behavioral analysis, and automated incident response. The system operates autonomously, providing edge devices with the equivalent of an immune system that continuously evolves against emerging attack vectors.

As AI workloads increasingly move to edge computing—where inferencing and even LLM fine-tuning occur locally—traditional network-based defenses often fall short due to latency constraints or intermittent connectivity. By embedding defenses directly into hardware and system calls, runtime security protects critical AI operations from model tampering, data poisoning, or adversarial attacks at the execution level, a capability institutional investors believe will be essential for edge AI adoption at scale.

Why are institutional investors and industry observers betting on embedded runtime security as a growth segment?

Investor interest is being driven by the convergence of two major trends. First, the shift toward edge processing is accelerating. Research firms project that by the end of 2025, roughly 75 percent of enterprise data will be generated and processed at or near the edge, up sharply from 10 percent in 2018. This move places greater risk on local devices that often lack robust security layers.

Second, AI’s increasing reliance on edge computing introduces new vulnerabilities. Analysts tracking AI security trends highlight threats such as adversarial sample injection, runtime model hijacking, and data exfiltration from GPU and NPU memory buffers. These attack vectors can only be effectively mitigated through protections embedded directly in the device rather than by delayed cloud detection.

European institutional sentiment is particularly strong, with investors viewing embedded runtime security as critical to data sovereignty and regulatory compliance. Analysts expect that companies offering hardware-level, AI-specific runtime protection will be positioned as strategic partners to manufacturers deploying sensitive or regulated AI systems.

How is Exein adapting its digital immune system for AI infrastructure and large language models?

The €70 million funding will accelerate Exein’s development of runtime security solutions tailored for AI accelerators, edge inferencing hardware, and on-device LLMs. According to company disclosures, its next-generation runtime agents will extend system call-level observability to neural network operations, offering the ability to detect abnormal parameter manipulation or unauthorized memory access during inference.

Exein has already partnered with hardware manufacturers such as Supermicro and Kontron to integrate its security stack into AI-optimized servers and industrial edge appliances. The strategy is to deliver zero-trust protections from boot sequence to inference, allowing real-time detection and response without requiring continuous cloud connectivity.

Future versions of the platform are expected to leverage embedded AI to autonomously adapt to evolving attack methods, making it suitable for autonomous vehicles, healthcare devices, and other mission-critical AI applications that demand immediate on-device response.

How do regulatory and market dynamics create a favorable environment for on-device runtime security?

Regulatory frameworks are pushing manufacturers toward embedded security by default. Europe’s Cyber Resilience Act, expected to take effect in 2026, and the U.S. Cyber Trust Mark mandate security certifications for connected products, with an emphasis on firmware-level protections. For edge AI deployments in healthcare, automotive, and industrial automation, runtime security ensures compliance while avoiding costly redesigns of legacy hardware.

The market is also shifting toward decentralization. Enterprises adopting edge AI for privacy, latency reduction, or sovereignty concerns are moving sensitive workloads out of centralized clouds, which increases their reliance on local device integrity. Institutional investors view these regulatory and market shifts as a structural tailwind for embedded runtime security vendors like Exein, whose technology can provide compliance and performance advantages simultaneously.

What do financial forecasts and investor sentiment indicate about the future of runtime security for AI infrastructure?

Exein’s rapid revenue growth and significant Series C valuation reflect rising investor conviction in on-device security as an emerging standard. With €20 million in expected revenues for 2025 and 450 percent year-over-year growth, analysts anticipate that embedded runtime security will transition from a niche IoT-focused solution to a mainstream requirement for AI infrastructure by the end of the decade.

The company is also preparing for strategic acquisitions to broaden its AI security capabilities, including potential expansion into hardware-based encryption and post-quantum secure boot technologies. Institutional investors suggest that if Exein maintains its current growth trajectory, it could position itself for a public listing as early as 2029 or 2030, setting a benchmark for embedded cybersecurity companies globally.

What challenges must be overcome for on-device runtime security to become the default for edge AI systems?

Despite strong momentum, adoption hurdles remain. Performance overhead in resource-constrained environments is a significant concern for developers, particularly when deploying AI inferencing models on specialized hardware like GPUs or NPUs. Achieving seamless integration without disrupting performance will be critical.

Standardization is another barrier. With diverse edge AI hardware ecosystems, vendors will need to collaborate closely with chip manufacturers, OEMs, and AI framework providers to create unified security layers. Finally, broader education about runtime-specific AI vulnerabilities is needed to drive adoption among enterprises still relying on perimeter-based defenses.

Can runtime security set the benchmark for AI infrastructure security at the edge in the next decade?

As enterprises scale AI workloads closer to the edge, embedded runtime security is emerging as a necessity rather than a luxury. Analysts project that by 2030, hardware-level runtime protection will be integrated into AI-capable devices as a standard feature, much like secure boot protocols today.

Exein’s funding, strategic partnerships, and regulatory alignment position it as one of the leaders shaping this transition. If it succeeds in extending its “digital immune system” from IoT to edge AI infrastructure, the company could play a pivotal role in defining how security is architected for the next generation of autonomous systems and critical infrastructure.