{"id":25564,"date":"2026-03-02T07:00:00","date_gmt":"2026-03-02T07:00:00","guid":{"rendered":"https:\/\/www.gsma.com\/newsroom\/?p=25564"},"modified":"2026-03-02T07:30:09","modified_gmt":"2026-03-02T07:30:09","slug":"gsma-launches-open-telco-ai-to-accelerate-development-of-telco%e2%80%91grade-ai","status":"publish","type":"post","link":"https:\/\/www.gsma.com\/newsroom\/press-release\/gsma-launches-open-telco-ai-to-accelerate-development-of-telco%e2%80%91grade-ai\/","title":{"rendered":"GSMA\u00a0launches Open Telco AI to\u00a0accelerate\u00a0development of\u00a0telco\u2011grade AI\u00a0\u00a0"},"content":{"rendered":"\n

New\u00a0initiative\u00a0is\u00a0supported\u00a0by\u00a0open-telco models,\u00a0including\u00a0a\u00a0new\u00a0family of\u00a0models\u00a0from\u00a0AT&T, compute from AMD\u00a0and\u00a0TensorWave,\u00a0datasets\u00a0from\u00a0researchers\u00a0and\u00a0a\u00a0new\u00a0portal\u00a0for industry\u00a0contribution and\u00a0collaboration\u00a0via\u00a0GSMA.com\/open-telco-ai<\/em><\/a>\u00a0<\/p>\n\n\n\n

2 March 2026<\/strong>, Barcelona \u2013 GSMA today launched Open Telco AI, a global industry initiative designed to accelerate telco-grade AI through open collaboration across operators, vendors, AI developers and academic institutions. The launch introduces a new portal for telco open models, data, compute and tools to accelerate the development and evaluation of telco-focused AI models, accessed via GSMA.com\/open-telco-ai<\/a>.   <\/p>\n\n\n\n

While frontier AI models have advanced rapidly, they continue to underperform on telecom specific tasks. 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 deployments1<\/sup> have been applied to network operations.  <\/p>\n\n\n\n

Open Telco AI meets this challenge by uniting industry and academic partners to build the foundations of telco\u2011grade AI models, data, compute, benchmarks and community. Progress is tracked through the Telco Capability Index, which measures model performance across an expanding set of telecom\u2011specific tasks. <\/p>\n\n\n\n

As founding supporters of Open Telco AI, AT&T and AMD are making significant contributions. AT&T is releasing a family of open telco-models developed and trained on open, publicly available data to be hardware and cloud\u2011agnostic, demonstrating that AI can deliver value across projects of any size and with varying levels of compute resources. AMD is providing compute capacity for model training, fine\u2011tuning, inference and evaluation through its GPU platforms, cloud partner TensorWave and open toolchains. <\/p>\n\n\n\n

The initiative is also supported by community programmes that bring together developers, researchers and operators to solve real-world telecom\u2011AI problems. This includes competitions such as the AI Telco Troubleshooting Challenge<\/a> which attracted over 1,000 registrations and will announce its winners at MWC26 Barcelona<\/a>. <\/p>\n\n\n\n

Louis Powell, Director of AI Initiatives, GSMA, said: \u201cToday\u2019s<\/em> 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.<\/em>\u201d <\/em> <\/p>\n\n\n\n

\u201cTelco networks are among the most demanding and regulated environments for AI and moving from promising demos to telco-grade performance requires an open foundation for data, workloads and compute,\u201d<\/em> said Philip Guido, executive vice president and chief commercial officer, AMD. \u201cThrough Open Telco AI, with GSMA and AT&T, AMD delivers the enterprise and AI compute needed to train, fine-tune and run open, telco-grade models efficiently from core to edge.\u201d<\/em> <\/p>\n\n\n\n

Andy Markus, Chief Data and AI Officer at AT&T, said: \u201cThe telecom industry needs AI that understands the realities of networks \u2013 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<\/em>\u2011<\/em>grade AI possible at scale. By contributing our expertise and shaping realistic test environments, we\u2019re 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\u2019s path toward intelligent, automated networks.<\/em>\u201d <\/p>\n\n\n\n

Building the Open Foundations of Telco-Grade AI<\/strong> <\/p>\n\n\n\n

The new portal will support the co\u2011creation of the essential building blocks for telco\u2011grade AI, including: <\/p>\n\n\n\n