Key Takeaways:
Nokia's commercial AI-RAN platform, built with Nvidia, promises to double spectrum capacity by 2028 and shift radio networks from hardware-defined to software-defined infrastructure.
Nokia on Wednesday launched the industry's first commercial AI-native radio access network platform, built on its anyRAN software and Nvidia's Aerial accelerated computing, threatening to upend the $40 billion RAN equipment market by decoupling performance upgrades from hardware replacement cycles.
"AI-RAN is the biggest innovation in radio in decades — it makes the network intelligent, extends AI into the physical world, and allows telcos to get more from their existing infrastructure, including a software upgrade path to 6G," Justin Hotard, president and chief executive officer at Nokia, said.
The platform has already demonstrated more than 20 percent spectral efficiency gains through AI-driven radio innovations, with Nokia targeting 50 percent improvement by 2027 and more than 100 percent by 2028 — effectively doubling the capacity of existing spectrum assets without new radio hardware. The system runs on Nvidia's Aerial AI-RAN platform using CUDA-accelerated baseband processing, and Nokia is offering three deployment paths: a GPU-powered AirScale capacity plug-in for existing Nokia customers, a standalone AI-RAN node for greenfield deployments, and cloud-native COTS server solutions through ecosystem partners. Pilot deployments begin by year-end, with commercial availability in 2027.
Nokia shares have surged about 140 percent over the past 12 months as the market re-rates the Finnish vendor from a slow-growth telecom equipment supplier to an AI infrastructure play. The stock trades at roughly 29 times forward non-GAAP earnings, with AI and cloud revenue growing 49 percent in the first quarter and the company booking about 1 billion euros in new AI orders. The question for investors is whether the revenue ramp can deliver operating leverage — group operating profit guidance of 2 billion to 2.5 billion euros remains unchanged despite the sharply higher growth assumptions.
How AI-RAN Changes the Economics of Radio Networks
Traditional RAN equipment relies on purpose-built hardware that must be physically replaced to improve capacity or add features. Nokia's AI-RAN platform shifts this to a software-defined model where spectral efficiency, energy consumption, and network automation improve through AI algorithms delivered via subscription. The company's new commercial model gives operators ongoing access to advanced AI features without hardware swaps, effectively turning the RAN into a continuously upgrading asset.
The competitive implications are significant. Ericsson, Nokia's primary rival, has been pushing its own "AI in RAN" software suite, announced earlier this year, which puts telco-grade AI models into existing baseband and radio equipment. But Nokia's platform is the first to integrate Nvidia's accelerated computing at the baseband level, creating a hardware-software stack that Jensen Huang, Nvidia's founder and chief executive officer, described as "transforming RAN into a planet-scale AI computer."
Nokia has already secured commercial validation. Taiwan Mobile signed a 5G expansion agreement that includes Nokia's AI-powered software for real-time automation, predictive analytics, and energy management. In South Korea, SK Telecom selected Nokia alongside Samsung, Ericsson, and HFR for the government's Hyper-AI Network Infrastructure pilot project, which will test AI-RAN for autonomous driving, humanoid robots, and patrol robotics.
The Supply Constraint and the Revenue Question
Nokia's challenge is no longer demand — it is supply. The company faces constraints across leading-edge semiconductors, memory, optical components, and indium phosphide production, limiting how fast it can ship products. Optical lead times currently run 12 to 18 months, with orders extending into 2027. Management has said it could ship more if it had access to more components and manufacturing capacity.
The supply bottleneck means Nokia's revenue ramp will be gradual rather than instantaneous. Network Infrastructure revenue grew 6 percent in the first quarter, while operating margin in the segment fell 30 basis points to 6.7 percent as the company invests in R&D, production capacity, and its new indium phosphide facility in San Jose. Nokia expects Optical Networks to reach a double-digit operating margin, and Infinera acquisition synergies are running ahead of schedule, but the consolidated margin story remains unproven.
For investors, the bull case hinges on whether Nokia can convert its order book — about 1 billion euros in AI and cloud orders in Q1 alone, compared with roughly 2.4 billion euros during all of 2025 — into sustained revenue growth and margin expansion. At 29 times forward earnings, the stock already prices in a successful transition. The second-quarter results, due July 23, will provide the first real test of whether the IP routing business is beginning to contribute meaningfully alongside the optical growth story.
This article is for informational purposes only and does not constitute investment advice.
