1stProtect and Multiverse Computing, a global leader in quantum-inspired artificial intelligence model compression, have announced a strategic alliance. The collaboration introduces an integrated solution for secure, fully localized AI inference at the network edge. The joint offering combines Multiverse’s capacity to drastically shrink advanced AI models for local execution with 1stProtect’s real-time, on-device enforcement protocols to monitor and govern autonomous agent behavior-enabling enterprises to run highly capable AI models without transmitting sensitive data to cloud environments.
As global enterprises race to transition generative and agentic AI systems from lab environments to active production, operational cost and administrative control have emerged as primary friction points. Cloud-reliant AI inference incurs significant bandwidth expenses while triggering complex data-sovereignty, privacy, and latency complications for heavily audited industries. Concurrently, the deployment of independent AI agents introduces novel runtime security vulnerabilities-including data leakage, prompt drift, and unauthorized execution paths-that legacy prompt filters and standard endpoint detection and response (EDR) platforms were fundamentally unequipped to neutralize. This partnership is engineered to solve these structural bottlenecks simultaneously.
Also Read: Cognizant and Domyn Partner to Deliver Sovereign AI Solutions for Regulated EMEA Enterprises
High-Fidelity Compression Meets Sub-100ms Runtime Enforcement
Under the terms of the agreement, the two organizations will co-develop an integrated technical architecture designed to support fully decentralized on-device AI deployments, requiring zero external cloud handshakes.
The unified software stack relies on two distinct technological pillars:
Multiverse‘s CompactifAI Optimization Platform: Utilizes quantum-inspired algorithms to compress the memory footprint of premier large language models (LLMs) by up to 95% while restricting accuracy loss to a mere 2% to 3%. This footprint reduction yields roughly 4x to 12x faster processing speeds and a 50% to 80% decrease in inference costs, allowing advanced models to operate smoothly on hardware form factors ranging from consumer smartphones to ruggedized industrial edge computers.
1stProtect’s Native Runtime Security Layer: Encapsulates the compressed on-device models within a continuous, low-overhead monitoring wrapper. The system delivers sub-100-millisecond threat interception capabilities, actively neutralizing prompt injection vectors, model extraction attempts, and unauthorized agent behavior the millisecond they materialize, whether the host device is operating online or entirely offline.
The consolidated solution is built specifically for high-security environments where cloud-based artificial intelligence has traditionally been unfeasible due to safety or infrastructure limitations. This includes air-gapped military networks, operational technology (OT) industrial systems, critical utility infrastructure, clinical healthcare equipment, and sovereign financial systems where intermittent connectivity or absolute data isolation is a strict operational requirement.
“This partnership brings a uniquely secure inference solution to the edge for the benefit of our customers. As enterprises move into an era of agentic and AI-driven threats, they need the freedom to run powerful models anywhere — without surrendering control. Together with Multiverse, we’re proving that on-device AI can be both remarkably efficient and rigorously protected. This partnership strengthens our commercial offering and validates the growing demand for runtime protection purpose-built for AI,” said Kervin Pillay, Chief Executive Officer, 1stProtect.
“1stProtect’s cutting-edge technology makes them the optimal partner to integrate with Multiverse’s on-device AI models. Their AI Agent Protection, a verified design together with Multiverse, is best in class. Compressing world-leading models so they fit on-device was the first breakthrough; making sure those agents run securely wherever they are deployed is what turns edge AI into an enterprise reality. This is the first killer application for on-device AI models and this partnership brings ubiquitous cyber security right to the device that matters: the one at your fingertips,” said Enrique Lizaso Olmos, Chief Executive Officer, Multiverse Computing.