Open RAN Development Company (ODC) has integrated its AI radio access network (AI RAN) software with Nvidia’s as part of the chipmaker's “all-American” 6G wireless initiative.

Demonstrated at Nvidia’s recent GTC DC event, the market-ready solution leverages the Nvidia AI Aerial platform, which combines accelerated computing platforms, software, and services for the management of wireless networks.

ODC claimed substantial gains compared to traditional CPU-based virtual RAN (vRAN) architectures, touting 40x faster physical layer (L1) signal processing, 7x increase in cell capacity, and 3.5x higher power efficiency per cell.

The Northern Virginia, U.S.-based firm also claimed 380x faster massive multiple-input, multiple output (MIMO) processing in the data link layer (L2) of a RAN setup.

The solution is set to launch in 2026, aimed at galvanizing 6G-ready deployments for AI inference use cases, with a network focus on real-time data sensing, analysis, and automated response.

“Our AI-RAN platform is not just about a faster more efficient network; it’s about creating an intelligent, city-scale compute fabric for all types of AI inference,” said Dr. Shaygan Kheradpir, chairman of ODC. “This enables operators to move beyond basic connectivity and enter the very large high-value market of delivering real-time intelligent applications for everything from enterprise AI to autonomous systems and public safety.”

“Our architecture provides the flexibility to run workloads on the optimal compute resource, whether CPU or GPU. By building on the Nvidia AI Aerial framework and leveraging the power of CUDA, we are not just accelerating the RAN; we are giving developers the tools to create a new generation of AI-driven applications that deliver unprecedented agility and performance for operators,” added Srinivas Gudladana, CEO of ODC.

With its framing as an all-American AI-RAN stack, ODC’s solution builds on Nvidia’s 6G AI-Wireless Intelligent Network (AI-WIN) venture. ODC joins the chip giant with Booz Allen, Cisco, MITRE, and T-Mobile on building America’s first AI-native wireless stack for 6G.

The project is underpinned by Nvidia AI Aerial, which is currently being piloted by SoftBank in Japan on various ventures.

In a recent blog, Nvidia developers gave a deeper look at the AI-WIN reference architecture, which hinges on a distributed User Plane Function (dUPF) built on AI Aerial.

The dUPF framework pushes data-plane intelligence from centralized data centers to the network edge, helping provide instantaneous response times for AI models by preventing transport delays in data between radio sites and distant core locations.

Nvidia is also growing its footprint in AI-RAN with a $1 billion investment in Europe's Nokia, as revealed last month.

Nokia said the investment will help its development of 5G and 6G RAN technologies. This includes a new AI-RAN solution centered around Nvidia's Aerial RAN Computer Pro (ARC-Pro) platform, with an expansion of Nokia's 5G and 6G RAN availability on Nvidia’s CUDA suite.