Using AI, DeepSig Advances Open, Intelligent Baseband RAN Architectures

Key Takeaways

• DeepSig removed pilot signal using AI
• OCUDU provides open-source carrier‑grade baseband stack
• U.S. DoD backs open RAN for defense resilience
• Ericsson and Nokia join OCUDU foundation, remain cautious
• AI‑RAN could capture one‑third market by 2029

Summary

DeepSig has leveraged deep‑learning models to eliminate the pilot signal in mobile networks, cutting signaling overhead without hurting performance. The startup is now a principal contributor to OCUDU, an open‑source carrier‑grade baseband stack launched by the Linux Foundation and the U.S. Department of Defense. OCUDU aims to democratize CU/DU software for 5G and early‑stage 6G, positioning smaller innovators alongside incumbents like Ericsson and Nokia. The partnership with Ireland‑based SRS marks DeepSig’s transition from niche research to a key player in the emerging AI‑native RAN ecosystem.

Pulse Analysis

Artificial intelligence is reshaping the radio access network by replacing handcrafted algorithms with data‑driven models. DeepSig’s breakthrough—eliminating the pilot signal—demonstrates how neural networks can trim signaling overhead, translating into modest latency gains and reduced spectrum waste. While the theoretical limits of Shannon’s law remain unchanged, even marginal efficiency improvements can generate billions in savings for cost‑sensitive operators as 5G densifies and 6G research accelerates.

The OCUDU initiative, co‑led by the Linux Foundation and the U.S. Department of Defense, seeks to codify a carrier‑grade, open‑source CU/DU stack. By exposing the baseband compute layer to community contributions, OCUDU lowers the cost of entry for startups and aligns with national security goals that demand transparent, auditable code for mission‑critical networks. DeepSig and Software Radio Systems were selected to deliver the first reference implementation, signaling a shift from proprietary silicon‑centric roadmaps toward modular, software‑defined architectures.

For the broader market, the move toward AI‑native RAN could reshape vendor dynamics. Analysts project AI‑RAN to represent roughly one‑third of the RAN market by 2029, driven by operator demand for automated optimization and the defense sector’s push for open, resilient infrastructure. Established players such as Ericsson and Nokia have joined the OCUDU ecosystem but remain cautious about fully abandoning their legacy stacks. The success of open‑source platforms will hinge on achieving carrier‑grade reliability and demonstrating clear cost‑benefit cases, setting the stage for a more competitive, innovation‑rich telecom landscape.