AI-RAN moves into enterprise infrastructure with joint SynaXG and Highway 9 platform

SynaXG and Highway 9 Networks have introduced a commercial AI-RAN deployment built on NVIDIA AI Aerial, combining private 5G connectivity with GPU-accelerated edge computing into a single enterprise platform. The system integrates radio access networks and AI workloads on shared infrastructure, targeting environments where connectivity and compute must operate as a unified layer.

The announcement centres on a cloud-native architecture that merges SynaXG’s AI-RAN platform with Highway 9’s AI Mobile Cloud. Together, the stack positions AI not as an overlay on existing networks, but as a workload that runs directly within the network fabric itself.

Converging RAN and AI workloads

At the core of the deployment is SynaXG’s software-defined AI-RAN platform, which supports both sub-6 GHz and millimetre wave spectrum. The system uses policy-based orchestration to allow AI workloads and network functions to run concurrently on shared infrastructure, removing the need for separate environments.

This design allows enterprises to treat radio access, compute, and data processing as programmable resources. It also introduces deterministic performance across mobile environments, with capabilities such as seamless mobility, high uplink capacity, and elastic scaling for dense or high-demand sites.

“Enterprises require a scalable platform that provides robust AI computing capabilities, along with a highly reliable and low-latency wireless network that is simple to deploy and future-proof,” stated Xin Huang, CEO of SynaXG. “SynaXG partnering with Highway 9 and NVIDIA deliver a high-performance solution to meet these needs.”

Private 5G meets edge AI control

Highway 9’s AI Mobile Cloud platform extends the system into enterprise IT environments, integrating private 5G, GPU-enabled edge computing, and orchestration within a unified control plane. The platform operates behind the enterprise firewall, enabling local breakout and policy-based workload placement across distributed edge nodes.

This approach allows enterprises to run on-premises LLM inference while keeping sensitive data within controlled environments. It also introduces multi-tenant resource guarantees, network slicing, and low-latency application placement, which are critical for environments requiring predictable performance and data sovereignty.

“AI is moving into the physical world — into factories, campuses and critical infrastructure,” said Allwyn Sequeira, founder and CEO of Highway 9 Networks. “Enterprises need an AI-ready mobile infrastructure layer that delivers deterministic performance, secure on-prem inference, and scalable edge compute. Together with SynaXG and NVIDIA, we are enabling enterprises to deploy AI workloads with ultra-reliability and the power of accelerated computing.”

GPU infrastructure as the common layer

The combined solution runs on a GPU-accelerated architecture, scaling from compact systems such as NVIDIA DGX Spark to larger deployments using NVIDIA Blackwell-based GPUs. This creates a shared compute layer that supports both network functions and AI inference workloads.

By consolidating these functions onto a single platform, the system enables applications such as autonomous robots, drones, real-time video analytics, and industrial automation to operate with low latency and consistent performance. It also provides a base for advanced capabilities including high-accuracy positioning, integrated sensing and communications, and distributed MIMO.

“AI-RAN is quickly becoming the blueprint for how telco operators and enterprises will build their next generation of networks,” said Soma Velayutham, Vice President of Telecoms and AI, NVIDIA. “With NVIDIA AI Aerial at the core, SynaXG and Highway 9 are showing how a single GPU-accelerated platform can deliver carrier-grade 5G, on-premise LLM inference, and secure edge AI services today – while creating a software-defined foundation that is ready for 6G.”

The companies plan to demonstrate the system at NVIDIA GTC 2026, with a live deployment highlighting how AI-RAN and cloud-based enterprise mobility can operate as a single programmable infrastructure layer.