Nokia Partners with NVIDIA and Orange to Develop AI-RAN Technology

2026-04-16

On April 15, 2026, Nokia and French telecommunications giant Orange jointly announced the expansion of their strategic partnership and collaboration with NVIDIA to co-develop and evaluate AI Radio Access Network (AI-RAN) technology. The core objective of this tripartite cooperation is to replace traditional CPUs with GPUs as baseband processing units, leveraging artificial intelligence algorithms to maximize spectral efficiency, significantly reduce energy consumption, and substantially enhance network capacity.

Under the cooperation framework, Nokia and Orange will collaboratively conduct identification, architectural design, and field evaluation of new AI-RAN capabilities. Key research directions include:

● Validating whether GPU-based wireless processing architectures can substantially improve link performance through optimized reception algorithms.

● Exploring the potential for new services such as environmental awareness and user behavior understanding after deeply embedding AI into the radio access network.

● Jointly developing spectral efficiency optimization solutions applicable to existing networks and future frequency bands—particularly the high-frequency 6GHz spectrum.

Technical Architecture and Platform Support

NVIDIA's AI Aerial platform provides critical technical support for this collaboration. The platform features software-defined accelerated computing capabilities, enabling flexible scheduling of 5G protocol stacks and AI workloads within a single GPU, thereby improving hardware resource utilization by two to three times. The platform's integrated Layer 1 and Layer 2 AI algorithms effectively enhance spectral efficiency, helping operators accommodate more users and deliver higher-performance services under identical resource conditions.

Orange Group's Chief Technology Officer noted that this tripartite collaboration facilitates a deeper understanding of how AI-native architectures can substantively improve the operational efficiency of core wireless algorithms. Nokia's Chief Technology and AI Officer stated that artificial intelligence is fundamentally transforming the design logic of communication networks, driving the deep integration of advanced AI capabilities with radio access network functions.

Industry Significance

This cooperation represents a further deepening of the strategic relationship between Nokia and NVIDIA. As early as October 2025, NVIDIA invested $1 billion in Nokia, becoming one of its major shareholders, and the two parties reached a strategic agreement to develop commercial-grade AI-RAN technology based on NVIDIA's platform.

Nokia has already launched the ARC-Pro accelerated computing platform based on NVIDIA's Grace CPU and Blackwell GPU, supporting software-upgrade evolution from 5G-Advanced toward 6G. This collaboration with Orange will provide critical scenarios for field validation of AI-RAN technology, laying a practical foundation for native AI wireless networks in the 6G era.

Semicone Electronic Chips Distributor

Semicone is a reliable supplier of semiconductor components, offering high-performance integrated circuits, active devices, passive devices, and power semiconductor products. Semicone's core competitiveness lies in stable supply, genuine product guarantee, and rapid procurement of components, which helps customers shorten project cycles and optimize supply chain costs. The following are today's selected popular IC models:

P9148NRGI8: High-performance multi-phase power supply controller, supporting intelligent current sharing and precise voltage regulation.
ATTINY461A-MUR: Atmel ultra-low power 8-bit AVR microcontroller, integrating 10-bit ADC and PWM module.
MCF52233CAF60: NXP ColdFire V2 core 32-bit microprocessor, built-in Ethernet MAC and encryption engine.
SI1062X-T1-GE3: Silicon Labs Sub-GHz wireless transceiver SoC, integrating 8051 core and multi-protocol stack.
MP5611GQT-Z: MPS highly integrated synchronous buck converter, supporting wide input voltage range and fast transient response.