Technical deep-dives on neural receivers, O-RAN architecture, and the path to inference silicon for the radio unit.
Barcelona gave the AI-RAN thesis its loudest endorsement yet. A dozen operators committed to AI-native 6G. Commercial hardware shipped. What MWC didn't answer is how deep the intelligence actually goes — and why the RU baseband wasn't on a single stage.
Read post →The industry has been running AI for the RAN in the cloud. 5.5G and 6G are about to make that model unworkable. The physics of higher frequencies demand denser cell sites, and denser cell sites demand that the intelligence moves to the edge.
Read post →The O-RAN Alliance standardized the fronthaul link between the DU and the RU — the protocols, the planes, the packet formats. But it left the RU's internal baseband processing as a vendor black box. For a company building neural inference inside the radio unit, that's a feature, not a bug.
Read post →The O-RAN 7.2x pipeline hasn't fundamentally changed since 3G. Signals pass through a rigid sequence of DSP blocks that were architected decades ago. It's time to rethink what happens inside the radio unit.
Read post →