# AI-RAN on NPUs: Baseband Processing Without Baseband Chips

arXiv: https://arxiv.org/abs/2607.04224 · HTML: https://mlsanity.com/paper/2607.04224/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 38 references that are in the map.
- First posted on arXiv 5 Jul 2026.

## Builds on (most cited first, 24 of 24)

- [In-Datacenter Performance Analysis of a Tensor Processing Unit](https://mlsanity.com/paper/1704.04760/) (2017)
- [An Introduction to Deep Learning for the Physical Layer](https://mlsanity.com/paper/1702.00832/) (2017)
- [Sionna: An Open-Source Library for Next-Generation Physical Layer Research](https://mlsanity.com/paper/2203.11854/) (2022)
- [Edge Artificial Intelligence for 6G: Vision, Enabling Technologies, and Applications](https://mlsanity.com/paper/2111.12444/) (2021)
- [NVIDIA Tensor Core Programmability, Performance & Precision](https://mlsanity.com/paper/1803.04014/) (2018)
- [Understanding O-RAN: Architecture, Interfaces, Algorithms, Security, and Research Challenges](https://mlsanity.com/paper/2202.01032/) (2022)
- [AI and ML Accelerator Survey and Trends](https://mlsanity.com/paper/2210.04055/) (2022)
- [A Survey on Deep Learning Hardware Accelerators for Heterogeneous HPC Platforms](https://mlsanity.com/paper/2306.15552/) (2023)
- [Toward Next Generation Open Radio Access Network--What O-RAN Can and Cannot Do!](https://mlsanity.com/paper/2111.13754/) (2021)
- [5G New Radio: Unveiling the Essentials of the Next Generation Wireless Access Technology](https://mlsanity.com/paper/1806.06898/) (2018)
- [AI-RAN: Transforming RAN with AI-driven Computing Infrastructure](https://mlsanity.com/paper/2501.09007/) (2025)
- [Design of a Standard-Compliant Real-Time Neural Receiver for 5G NR](https://mlsanity.com/paper/2409.02912/) (2024)
- [Beyond Connectivity: An Open Architecture for AI-RAN Convergence in 6G](https://mlsanity.com/paper/2507.06911/) (2025)
- [Benchmarking Edge AI Platforms for High-Performance ML Inference](https://mlsanity.com/paper/2409.14803/) (2024)
- [X5G: An Open, Programmable, Multi-vendor, End-to-end, Private 5G O-RAN Testbed with NVIDIA ARC and OpenAirInterface](https://mlsanity.com/paper/2406.15935/) (2024)
- [Hardware Acceleration for Neural Networks: A Comprehensive Survey](https://mlsanity.com/paper/2512.23914/) (2025)
- [Hardware Acceleration for Open Radio Access Networks: A Contemporary Overview](https://mlsanity.com/paper/2305.09588/) (2023)
- [Towards AI-Driven RANs for 6G and Beyond: Architectural Advancements and Future Horizons](https://mlsanity.com/paper/2506.16070/) (2025)
- [ML-Enabled Open RAN: A Comprehensive Survey of Architectures, Challenges, and Opportunities](https://mlsanity.com/paper/2604.01239/) (2026)
- [From Centralized RAN to Open RAN: A Survey on the Evolution of Distributed Antenna Systems](https://mlsanity.com/paper/2411.12166/) (2024)
- [A Compute&Memory Efficient Model-Driven Neural 5G Receiver for Edge AI-assisted RAN](https://mlsanity.com/paper/2508.12892/) (2025)
- [Debunking Seven Myths about 5G New Radio](https://mlsanity.com/paper/1908.06152/) (2019)
- [Efficient Parallelization of 5G-PUSCH on a Scalable RISC-V Many-core Processor](https://mlsanity.com/paper/2210.09196/) (2022)
- [TensorPool: A 3D-Stacked 8.4TFLOPS/4.3W Many-Core Domain-Specific Processor for AI-Native Radio Access Networks](https://mlsanity.com/paper/2604.02291/) (2026)

## Abstract (arXiv)

AI-RAN aims to unify artificial intelligence and radio access network workloads on a shared compute substrate. While this paradigm has so far been demonstrated primarily on Graphics Processing Units (GPUs), it remains unclear whether Neural Processing Units (NPUs), which are AI accelerators optimized for inference, can also support wireless baseband processing. Here, we provide the first affirmative answer by resolving the fundamental mismatch between baseband workloads and NPU architecture. A computational isomorphism exists: matrix and vector engines NPUs dedicate to inference inherently cover physical-layer operations. Yet NPU architectures are natively shaped for dense-tensor AI inference, not baseband. This architectural mismatch surfaces as opposing optimization objectives: traditional baseband minimizes arithmetic operations, whereas NPU performance demands maximizing engine utilization. We close this gap by reconstructing communication algorithms onto AI compute primitives, prioritizing engine utilization over arithmetic count. We validate this with a complete OFDM transceiver on an Ascend 310B1 edge NPU, demonstrating end-to-end over-the-air transmission via USRP X300 at 3.0 GHz.