MWC26: GSMA launches Open Telco AI to drive telco-grade AI models

• ‘AI does not yet speak telco’, according to the GSMA
• Its new initiative brings together industry and academia to build AI models specifically for telecom operators
• AT&T, AMD and neocloud TensorWave are making significant early contributions

BARCELONA – #MWC26 – The GSMA has announced a new initiative, Open Telco AI, that is designed to accelerate the availability of suitable telco-grade AI through open collaboration across operators, vendors, AI developers and academic institutions.

The move highlights a key current predicament facing telcos that are adopting AI-native strategies: Generic frontier AI models aren’t tuned for the specific needs of network operators.

Louis Powell, director of AI initiatives at the GSMA, stated: “Today’s AI models still fall short of the complexity, precision and reliability the telecom industry demands. Put simply, AI does not yet speak telco and operators are often deploying technology that cannot meet the required levels of accuracy, safety or efficiency. Establishing clear benchmarks and collaborating across the industry on datasets, models and agentic systems is essential. Open Telco AI provides a shared foundation designed to close this gap, an approach that other regulated sectors, such as finance and healthcare, can follow.” 

Open Telco AI introduces a new portal for AI models, data, compute and tools that aims to aid the development of telco-focused AI models.

The GSMA says that frontier AI models are continuing to underperform when it comes to telecom-specific tasks, noting that only 16% of telecom generative AI (GenAI) deployments have been applied to network operations, based on figures from GSMA Intelligence’s Telco AI report.

It claims its new initiative will help its members overcome this challenge by pairing industry with academia to build new telco-grade AI models and benchmarks, which will be tracked through the Telco Capability Index that has been developed to measure model performance across an expanding set of telecom‑specific tasks. 

The initiative benefits from a number of big-name supporters. 

AT&T is contributing to Open Telco AI by “releasing a family of open telco models developed and trained on open, publicly available data to be hardware and cloud‑agnostic, demonstrating that AI can deliver value across projects of any size and with varying levels of compute resources,” noted the GSMA. 

Andy Markus, chief data and AI officer at AT&T, stated: “The telecom industry needs AI that understands the realities of networks – not only generic models repurposed for telco tasks. Through Open Telco AI, AT&T is helping build the datasets, models and evaluation frameworks that make telco‑grade AI possible at scale. By contributing our expertise and shaping realistic test environments, we’re demonstrating how generative and agentic AI can improve customer experience, reduce operational friction and ultimately create new value. This collaboration with GSMA is accelerating the industry’s path toward intelligent, automated networks.” 

Meanwhile, chip giant AMD is providing compute capacity for model training, fine‑tuning, inference and evaluation through its GPU platforms and open toolchains and by enlisting the support of neocloud TensorWave, which operates one of the world’s largest all-AMD GPU clouds.

Other founding partners contributing data, models and use cases include Huawei, KDDI, KPN, LG Uplus, Nvidia, Ooredoo, Orange, SK Telecom, Softbank, Swisscom, Turkcell and more, alongside academic institutions from the US, Europe and the Middle East.

- James Pearce, Editor, TelecomTV