Telecom Giants Launch Open-Source AI Initiative
The GSMA, backed by industry heavyweights AT&T and AMD, has launched “Open Telco AI.” The global initiative aims to accelerate the development of telecom-grade artificial intelligence by creating open-source models and standardized datasets for the industry.
The initiative was born from a significant performance gap; general-purpose AI models often struggle to interpret complex network data and correctly understand telecom standards documentation. This has limited the application of generative AI in network operations, with one report indicating only 16% of such deployments have been focused on the core network. AT&T is contributing by releasing a family of open-source "telco-models" trained on publicly available data, which are designed to be independent of specific hardware or cloud platforms. AMD, along with its cloud partner TensorWave, is providing the essential high-performance GPU computing power needed for training, fine-tuning, and evaluating these specialized AI models. This alliance extends far beyond its founding partners, gathering a broad coalition of operators, tech firms, and academic institutions. Contributing partners include notable names like Huawei, KDDI, Nvidia, Orange, SK Telecom, Swisscom, Google Cloud, IBM, Telefónica, and Vodafone, among many others. A key component of the Open Telco AI initiative is a new online portal that will serve as a hub for collaboration. It will provide access to telco-specific models, a library of data including knowledge graphs and curated standards, and tools for benchmarking model performance on telecom tasks. This move is part of a wider industry trend to tailor AI for network functions. It follows the establishment of the AI-RAN Alliance in 2024, which includes members like Nokia, Samsung, and T-Mobile, and focuses on integrating AI with Radio Access Network (RAN) technology to improve efficiency for 5G and 6G.
