According to GSMA Intelligence, telecom operators in 2024 primarily deployed AI to cut costs through low-risk use cases such as customer care automation, network optimization, and predictive maintenance.
However, there is a growing trend toward AI deployments aimed at generating new revenue, particularly in the US (20 percent of deployments), China, and other parts of Asia (25 percent), where operators are experimenting with GPU-as-a-Service (GPUaaS), AI inference, and AI factories — led by companies like Verizon and SK Telecom. The commercial outcomes of these initiatives, expected in late 2025 and 2026, will indicate the success of telco AI monetization.
A key emerging opportunity lies in sovereign AI, where operators retain and process AI workloads domestically to ensure data control and compliance. Examples include collaborations by Ooredoo Qatar with Ericsson, Indosat, and Bell Canada with Cohere. This approach strengthens telcos’ strategic advantage over hyperscalers due to their national infrastructure role.
On the infrastructure front, Nvidia continues to dominate the GPU market for AI workloads, far ahead of AMD and Intel. While competition may increase, Nvidia remains the leader unless major geopolitical shifts favor Chinese rivals. In cloud partnerships, Google leads in large telco collaborations, while Microsoft holds a strong position with small and mid-sized operators.
GSMA Intelligence data shows that customer care remains the leading focus area for telco AI, accounting for 47 percent of all deployments. Operators use AI in call centers and sales channels to reduce costs, improve upselling and cross-selling, and curb customer churn.
About 70 percent of AI projects are already live, integrated into daily operations, while 30 percent remain in trial or planning stages. Unlike traditional network rollouts, AI adoption will follow a cyclical pattern, with continuous testing and iterative deployment rather than a single linear transition.
Currently, around 20 percent of telco AI deployments have a direct revenue objective, marking a gradual shift from cost-saving initiatives to monetization efforts such as GPU-as-a-Service (GPUaaS) and agentic AI models. This trend is expected to accelerate through 2025 and 2026 as trials mature into commercial offerings.
Larger telecom operators maintain a 20 percent “size premium” in AI adoption compared to smaller peers, driven by bigger budgets, stronger R&D capabilities, and first-mover ambitions — though some smaller players are emerging as exceptions to this pattern.
Optus launched Optus Expert AI, an agentic AI solution designed to support its frontline teams, improving customer interaction efficiency and operational intelligence.
Indosat introduced Vision AI, an AI-based smart surveillance solution that enhances network and site security through intelligent video analytics.
Korea Telecom unveiled the KT AI Station, a mobile AI experience centre aimed at showcasing and promoting hands-on AI applications for consumers and businesses.
True Corporation deployed its AI-CODC system, which uses AI to monitor and manage base station signals, improving network performance, reliability, and operational stability.
Ooredoo Qatar launched a sovereign AI cloud powered by Nvidia Hopper GPUs, enabling secure, localized AI compute capabilities that align with national data sovereignty goals.
Umniah partnered with Ericsson to implement AI/ML-based network energy optimization solutions, reducing power consumption and supporting sustainability initiatives.
AT&T, working with Aira Technology and Ericsson, successfully tested an AI-generated rApp on Ericsson’s Intelligent Automation Platform, marking progress in autonomous network management.
Bell Canada and Cohere formed a strategic partnership to deliver sovereign AI-powered solutions tailored for government and enterprise clients, enhancing data control and compliance.
Etisalat e& launched an AI-powered drone solution for tower inspections, improving maintenance efficiency and safety while reducing operational costs.
Elisa expanded its collaboration with Google Cloud to develop an AI-driven autonomous network, advancing automation and real-time network optimization capabilities.
South Korea’s Ministry of Science and ICT selected five elite teams to drive the Sovereign AI Foundation Model project, strengthening the country’s national AI ecosystem.
UK telecom regulator Ofcom released a strategy for safe and innovative AI use across telecoms, broadcasting, postal, and online platforms, focusing on transparency and accountability.
Brazil’s telecom regulator Anatel launched a public consultation on the use of AI in telecom networks and cybersecurity, aiming to establish frameworks for responsible AI integration.
Baburajan Kizhakedath