#USA #New York #AI-RAN #Cerberus Capital #NVIDIA Aerial

Revolutionizing Wireless Networks with AI Technology

In a groundbreaking move within the telecommunications industry, Cerberus ODC has teamed up with NVIDIA to launch an innovative AI-native Radio Access Network (RAN) stack. This remarkable development aims to enhance 5G capabilities while paving the way for the transition to 6G technologies.

Key Developments

On October 30, 2025, Cerberus ODC, a subsidiary of Cerberus Capital Management, proudly announced the successful integration of its advanced AI-driven RAN software with NVIDIA's AI Aerial platform. This effort culminates in the creation of the first fully functional All-American AI-RAN stack, designed to achieve unprecedented performance metrics in wireless networking. The new capabilities were showcased at the prestigious NVIDIA GTC DC conference held from October 27 to 29.

Performance Breakthroughs

Harnessing the powerful NVIDIA AI Aerial platform, which encompasses the innovative Grace Hopper Superchip alongside CUDA-accelerated libraries, the AI-native RAN software stack reported remarkable performance upgrades when compared to traditional architectures reliant on CPUs. Significant advancements include:

• - 380x faster Massive MIMO Layer-2 processing
• - 40x quicker Layer-1 signal processing
• - A remarkable 7-fold increase in cell capacity
• - 3.5x enhancement in power efficiency per cell

These 'x-factor' improvements underscore the potential of an architecture that delegates intensive computational tasks to the GPU, fundamentally transforming the operational efficiency and capacity of wireless networks.

Establishing a Distributed AI Engine

Beyond merely upgrading RAN performance, the collaboration between Cerberus ODC and NVIDIA has enabled the transformation of network infrastructures into comprehensive, distributed AI engines tailored for inference workloads. The advancements in performance and latency provided by the AI-RAN architecture bring forth new opportunities for the deployment of complex AI applications that necessitate real-time sensing, analysis, and action directly executed from the network.

According to Dr. Shaygan Kheradpir, Chairman of ODC, this milestone signifies an essential evolution toward the establishment of AI-native wireless infrastructures. "Our AI-RAN platform goes beyond improving network speed and efficiency; it lays the foundation for a city-scale intelligent compute fabric for various AI inference applications. This will allow operators to transition from basic connectivity services into the lucrative market for advanced real-time intelligent applications, spanning from enterprise AI applications to autonomous systems and public safety solutions," he stated.

The Advantage of Software-Defined Architecture

The seamless integration achieved from ODC's Layer 2/Layer 3 software to NVIDIA's Layer 1, alongside commercial 5G core technologies and devices, illustrates the strength of a software-defined architecture. This framework delivers exceptional agility, adapting the efficiencies derived from cloud technologies into the telecommunications landscape, effectively expediting the transition from prototype to production phases from years to mere days. Additional features and AI capabilities can be continuously deployed and updated even while the network remains operational.

Srinivas Gudladana, CEO of ODC, emphasized that AI-RAN marks a transformative shift in network operation paradigms. His insight highlighted the flexibility of the architecture, allowing workloads to be processed on the most optimal computing resources, whether CPU or GPU. "By leveraging the NVIDIA AI Aerial framework and CUDA’s capabilities, we are not just advancing RAN efficiency; we are equipping developers with the required tools to innovate an entirely new class of AI-centric applications delivering unparalleled performance for operators," he stated.

Preparing for Real-World Deployment

The AI-RAN solution has attained full production maturity status and is poised for deployment in real-world scenarios. With this milestone, the Nvidia-ODC partnership has effectively positioned the AI-RAN to surpass existing legacy RANs, propelled by American innovation. Operators can now leverage 5G Advanced-ready applications for a vast range of AI inference workloads, from general enterprise needs to highly specialized 'AI for the Physical World' operations.

Starting in the first quarter of 2026, operators will be able to roll out this next-generation AI-RAN solution, which has been meticulously crafted for various mission-oriented enterprises. To explore the technical intricacies of the system further, readers are encouraged to visit ODC's tech blog.

About ODC

Open RAN Development Company (ODC) stands at the forefront of technological innovation, providing AI-native, software-defined RAN solutions that cater efficiently to diverse compute environments. ODC's platform is designed to empower network operators, facilitating the swift adoption of AI-RAN infrastructure while unlocking fresh revenue opportunities amid the growing demand for physical-world AI applications. For more information about ODC, please visit www.orandevco.com.