- General-purpose LLMs struggle to interpret network data, understand standards documentation and automate network operations
- The Open Telco AI initiative seeks to enable improved AI telco suitability
- AMD and AT&T are founding supporters, along with more than two dozen organizations worldwide
Frontier AI models are great general-purpose tools, but they fall short on telco-specific tasks. GSMA is building a coalition to change that, helping tailor AI models specifically for telco requirements.
At Mobile World Congress 2026 in Barcelona today, GSMA launched Open Telco AI, with founding supporters AT&T and AMD. The initiative is a global alliance of operators, vendors, AI developers and academic institutions to collaborate on accelerating AI, developing models, shared data, computer benchmarks and community.
"Many general-purpose models struggle to interpret network data, understand standards documentation or automate network operations with sufficient accuracy. This performance gap limits progress: only 16% of telecoms GenAI deployments have been applied to network operations," GSMA said in a statement.
AT&T and AMD contributions
As part of the initiative, AT&T is releasing a family of open telco models developed and trained on open, publicly available data to be hardware and cloud-agnostic. AMD is providing compute capacity for model training, fine-tuning, inference and evaluation through its GPU platforms, cloud partner TensorWave and open toolchains.
The initiative will develop high performance open weight models for telecom tasks, from network troubleshooting to standards interpretation, including models of multiple sizes and architectures from AT&T, a radio-frequency language model from Khalifa University called RFGPT and a Large Telco Model (LTM) from AdaptKey AI built on Nvidia Nemotron.
For open data, Open Telco AI uses a library of knowlege graphs, embeddings and fine-tuning datasets of text, logs and curated standards material from GSMA, Huawei Technologies France, Khalifa University, Mantis NLP, NetoAI, Pleias, Purdue University, The University of Texas at Dallas, University of Leeds and Yale University and pipelines for generating synthetic data from NVIDIA. Open Telco AI has access to compute and open toolchain for projects training and inferencing open models via AMD and TensorWave.
And for benchmarking, a leaderboard assesses model performance.
In addition to AMD and AT&T, the initiative is supported by more than two dozen organizations worldwide.
What other industry players are doing with network AI
GSMA joins multiple efforts to customize AI for telco needs. Nvidia said in March 2025 it is developing LLMs specialized for telcos — large telco models (LTMs) — and AI agents that can configure networks in minutes, where operations formerly took days. Network automation emerged as the top use case for telecom AI ROI in a Nvidia survey released in February.
In June, Cisco debuted AgenticOps, an AI-driven approach to managing IT operations without human oversight, built on an advanced networking LLM called the Deep Network Model.
Read all of our coverage from Mobile World Congress 2026 in Barcelona here.