DeepSig becomes a founding member of the Linux Foundation’s OCUDU initiative, moving a government-backed CU/DU stack into open source governance to speed interoperable, AI-native 5G and early 6G deployments.
DeepSig has joined the OCUDU Ecosystem Foundation as a founding member, backing a vendor-neutral, open-source framework under Linux Foundation governance to standardise interoperable Centralised Unit and Distributed Unit software for Open RAN.
OCUDU establishes an open reference platform, integration tooling and continuous validation environments designed to cut fragmentation and enable production-grade deployments across 5G and early 6G networks. The aim is to make open source the default foundation for next-generation RAN infrastructure rather than a research effort.
The stack builds on the Software Radio Systems 5G CU/DU solution, previously selected by the US FutureG Office with the National Spectrum Consortium, and has now transitioned into a fully open-source project hosted by Linux Foundation. A three-year funded programme supports a DeepSig–SRS partnership to expand capabilities spanning advanced 5G features, early 6G, AI-native RAN and accelerated compute. An initial release is already live.
DeepSig contributes AI-RAN expertise and hardware acceleration support, including NVIDIA GPUs, enabling spectrum sensing, neural receivers, neural scheduling and AI-native air interfaces, alongside efficient RAN processing and reference designs.
“OCUDU creates the foundation for a more open and software-driven RAN ecosystem,” said Jim Shea, CEO, DeepSig, noting that an open CU/DU platform enables rapid experimentation, interoperability and deployment.
Arpit Joshipura, General Manager, Networking, Edge and IoT, Linux Foundation, added that DeepSig’s participation helps “advance interoperable CU/DU software and speed innovation for 5G and early 6G networks.”
The foundation now shifts from government-backed development to neutral governance, inviting operators, vendors, cloud providers and researchers to collaborate at scale.
