Orange to test Nokia AI-RAN in live network, starting small
Orange will deploy Nokia's AI-RAN platform in its live network, testing performance, energy efficiency, and total cost of ownership. The operator will consider all three deployment models, starting with plug-ins for existing AirScale base. Orange also sees potential for AI applications beyond RAN optimization, but the economics of shared compute remain unproven.
Orange's trial begins modestly, evaluating radio performance, energy use, and operational impact in real conditions. The operator will examine all three Nokia configurations, starting with plug-ins for existing AirScale equipment. Nokia's lab tests show roughly 20 percent spectral efficiency gains, targeting 50 percent by mid-2027 and over 100 percent in 2028.
Orange joins carriers including T-Mobile, NTT Docomo, Softbank, and several Middle Eastern operators exploring this technology. Orange's CTO noted future AI-driven traffic may demand more uplink capacity and lower latency. The company sees potential for edge AI applications like sensing and positioning, though the business case for shared compute remains uncertain.
If AI-RAN efficiency gains hold in live conditions, operators could reduce energy costs and extend network capacity without major hardware overhauls. This may benefit consumers through more reliable service and potentially lower prices, while enabling new edge-AI applications in sensing and positioning. However, the economics remain unproven, and smaller carriers may struggle to justify the investment, potentially widening the gap between large and small operators in network capability and AI-driven services.