Telecom operators around the world are poised to earn over $21 billion in revenue by 2030 through the provision of GPU-as-a-Service (GPUaaS), as demand for artificial intelligence (AI) capabilities surges globally. A new report from ABI Research highlights that telecoms are uniquely positioned to support national and regional AI infrastructure, especially in response to rising concerns over AI sovereignty—where governments aim to maintain control over their AI systems and data.
ABI Research analyst Larbi Belkhit emphasized that while the telco GPUaaS market is still in its infancy, the momentum is building fast. Telecom companies are increasingly transforming their data centers into “AI factories” capable of supporting heavy AI workloads. This infrastructure shift is being driven by the expanding enterprise use of AI and governmental push for localized AI processing to meet sovereign data and regulatory demands.
By the first quarter of 2025, at least 17 telecom companies had already built or announced plans for NVIDIA-powered AI factories. Prominent names include SoftBank, Reliance Jio, TELUS, Telenor, and Ooredoo. These facilities are not only powering internal AI applications but are also positioned to offer GPUaaS to external clients, creating a new revenue stream.
The Asia-Pacific region currently leads in the deployment of these AI factories, followed by Europe, the Middle East, and Africa. ABI forecasts that 2027 will mark a major turning point, where GPUaaS revenues for telecoms are expected to rise sharply as the market matures and more national AI infrastructure comes online.
Ultimately, the move into GPUaaS and broader AI-as-a-Service (AIaaS) models could provide telecom operators with a strategic advantage over traditional cloud providers. By leveraging their existing infrastructure and regulatory familiarity, telcos could become the trusted backbone of sovereign AI, tapping into sensitive but high-value sectors such as finance, healthcare, and government services.
Source: The guardian