电信人工智能：2025年第一季度市场状况

AI and networks The GSMA is a global organisation unifying the mobile ecosystem todiscover, develop and deliver innovation foundational to positive businessenvironments and societal change. Our vision is to unlock the full powerof connectivity so that people, industry, and society thrive. Representingmobile operators and organisations across the mobile ecosystem andadjacent industries, the GSMA delivers for its members across three GSMA Intelligence is the definitive source of global mobile operator data,analysis and forecasts, and publisher of authoritative industry reports andresearch. Our data covers every operator group, network and MVNO inevery country worldwide – from Afghanistan to Zimbabwe. It is the most GSMA Intelligence is relied on by leading operators, vendors, regulators,financial institutions and third-party industry players, to support strategicdecision-making and long-term investment planning. The data is used as We invite you to find out more at www.gsma.com Our team of analysts and experts produce regular thought-leadingresearch reports across a range of industry topics. www.gsmaintelligence.cominfo@gsmaintelligence.com Authors Peter JarichHead of GSMAIntelligence Sayali BoroleSenior Analyst Tim HattHead of Research AI for networksversus networksfor AI When we began our State of the Market series in 2024, the goalwas to provide a view of the many ways operators were innovatingwith AI, and flag some dynamics that require more attention if The series has covered a lot of ground, including a deep dive into openversus proprietary large language models (LLMs), the role of energypressures in scaling the use of AI, and a framework for how to think This latest iteration includes one of the most important industry dynamics:AI for networks versus networks for AI. AI has been used by telecoms operators in their network operationsfor years – whether to guard against threats, minimise energy usage oroptimise RAN performance. As AI’s use proliferates, operators are in aunique position to enable its growth through network investment andevolution. Providing a foundation from which AI can thrive will require Peter JarichHead of GSMA Intelligence About the series This is the final report in a four-part series on AI strategy in telecoms. This edition examines the impactof AI on telecoms networks, and the flipside of how network capabilities can be leveraged for AI. In tandem with the report series,GSMA Intelligence has developed anAI benchmark to track AI use and Market context Deep dives Leader profiles •A recap of criticalindustry developmentsand implications Deep dive:AI for networksPage 1604Deep dive:networks for AIPage 2105Market contextPage 802Leader profiles:telco AI in actionPage 1103Executive summaryPage 501 01 Executivesummary Summary AI-driven network efficiency and cost optimisation Networks for AI: inferencing and edge 01 Smarter, more agile network infrastructure is essential tosupport AI applications. AI requires significant processing andcomputing power. Cloud integration enables AI training andlarge-scale computing, while edge computing reduces latencyand operational costs by processing data closer to its source. AI is making networks intelligent, efficient and cost effective.Use of AI has evolved from automating tasks to optimisinglarger systems. In network operations, predictive maintenance,self-healing capabilities and intelligent traffic management havereduced downtime and improved cost efficiencies. Taking a AI for networks: managing trafficand bandwidth demand 05 Networks for AI: new revenue models 02 AI inference opens up new use cases in areas such asindustrial robotics, digital twins, IoT and VR/AR, allowingoperators to create new revenue streams with enterprisecustomers. GPU-as-a-service, AI model licensing andlocalised AI infrastructure all offer monetisation opportunitiesfor operators. Local AI processing is also seeing growing Networks are the core of the operator business. With 5Gadoption and AI-powered applications driving exponentialtraffic growth, networks must handle increasing bandwidthdemand. AI solutions can help networks keep up through AI for networks: energy and sustainability 03 The RAN is the most energy-intensive part of telecomsinfrastructure, at 80–85% of total consumption – a costline that has remained stubbornly high. AI-powered RANIntelligent Controllers and traffic steering can enhanceresource allocation, automate fault detection and support AI in numbers 62% 53% 80% 3× AI for maintenanceand planning AI investmentimplications Expertise andexperience paramount Growth imperative Improving customerexperience and supportingnew services beat opexand capex savings as the Across many AI investmentrequirements, 63% ofoperators prioritised AIinvestment in upskilling,followed closely by networkcapacity (62%). The formerpoints to the requirement 15% of operators expectadded data traffic to be thebiggest network impac