NVIDIA and Telecom Partners Push Open, AI-Native 6G Networks

NVIDIA and a coalition of major telecom operators and tech vendors have announced plans to collaborate on building next-generation 6G wireless networks using open, AI–native platforms, a move that could reshape the architecture of future mobile infrastructure.

The initiative, unveiled during Mobile World Congress in Barcelona, brings together companies including BT Group, Cisco, Deutsche Telekom, Ericsson, Nokia, SK Telecom, SoftBank and T-Mobile, along with consulting firm Booz Allen Hamilton and research organization MITRE. The group said it will focus on developing open, software-driven wireless platforms that embed AI throughout the network stack.

AI and Telecom

The effort reflects a growing belief within the telecom sector that future networks must be built differently from earlier generations. While 5G expanded bandwidth and capacity, it was not designed with AI-driven services in mind. The next generation of wireless technology is expected to support far more complex workloads, including networks of autonomous machines and sensors interacting in real time.

AI, of course, will be central to that transformation. By integrating AI directly into the radio access network, edge infrastructure and core systems, telecom operators could operate networks that continuously adapt through software updates rather than hardware upgrades.

The approach is built around AI-RAN (AI-Radio Access Network), a software-defined radio access architecture that allows network processing and AI workloads to run on the same accelerated computing platforms. According to NVIDIA, this design would enable telecom providers to manage network traffic more dynamically while supporting new AI applications at the edge of the network.

For NVIDIA, the initiative represents a strategic effort to extend its role in AI infrastructure into the telecom sector. The chip giant has already participated in several research initiatives focused on 6G development across the US, Europe and Asia.

As wireless networks evolve into platforms for intelligent services, telecom operators are likely to face intense demand for computing power and flexible software architectures. The companies participating in the coalition say the goal is to establish the foundation now (and presumably their leadership in that foundation), before the next generation of wireless reaches commercial deployment.

Growing Interest in AI-RAN

Industry interest in AI-RAN is already moving beyond laboratory testing. Several telecom operators have conducted early field deployments using NVIDIA-based platforms. T-Mobile recently demonstrated a system capable of running both AI applications and radio network functions simultaneously on the same infrastructure, supporting services like video streaming and AI-assisted captioning over a live 5G connection.

SoftBank reported a separate trial in which a software-defined network running on NVIDIA hardware achieved a multi-layer massive MIMO (Multiple-Input, Multiple-Output) configuration, a technical milestone considered important for future high-capacity networks.

Beyond these demonstrations, the broader industry ecosystem around AI-RAN continues to expand. Equipment vendors and server manufacturers have begun introducing systems designed to support AI-driven radio networks. These include hardware platforms that combine general-purpose processors with graphics processing units to handle both networking and AI computation.

Support for Open Source Networking

The coalition’s emphasis on open platforms also reflects a shift in telecom policy discussions. Some governments, including the US, have expressed support for open-source networking architectures as a way to reduce dependence on a small number of infrastructure vendors and to encourage innovation across the telecom sector.

Advocates argue that open software platforms could allow developers and smaller tech firms to contribute algorithms and applications directly into network systems, accelerating experimentation. Critics, however, note that hardware dependencies, particularly the reliance on specialized computing platforms, may still shape how open these systems ultimately become.

Even so, the momentum behind AI-native networks is growing. Industry surveys indicate that telecom operators are increasingly investing in AI-driven network architectures and exploring ways to automate network management through machine learning models.

Those developments could alter how mobile networks are built and operated in the coming decade. Commercial deployment of 6G is widely expected to begin around 2030, with early trials anticipated before the end of the decade.