Telecom operators globally have a large opportunity to expand their business beyond connectivity and become the backbone of AI infrastructure for national or regional AI demand.

According to global technology intelligence firm, ABI Research, the telco industry is set to generate over $21 billion in GPUaaS revenue by 2030. This revenue growth will be driven by the continued growth in demand for GPU compute by enterprises and the growing requirement of AI sovereignty by governments.

Industry Analyst at ABI Research, Larbi Belkhit, said: “While the telco GPUaaS market is still at an incredibly early stage, there is significant momentum behind the buildout of AI factories and upgrading existing data centre portfolios to better support AI workloads.
 
“Demand for access to GPU compute continues to grow, but alongside that is a growing necessity for AI sovereignty, and telcos are extremely well positioned to be the preferred and trusted partner for national or regional AI inferencing workloads.”
 
As of Q1 2025, 17 telcos have announced or built NVIDIA AI factories, which can be utilised not just for internal workloads, but GPUaaS offerings too. These include SoftBank, Reliance Jio, Telenor, TELUS, Ooredoo, and others.
 
The majority of the announced NVIDIA AI factories are in the Asia-Pacific region thus far, followed by Europe and the Middle East and Africa markets. As the buildout continues and telcos become more mature in distributed GPU compute offerings, 2027 is expected to be the inflexion point for when GPUaaS revenue for telcos begins to rapidly grow.
 
Belkhit added: “While highly regulated and sovereign-sensitive verticals are prime opportunities, GPUaaS is merely a gateway to a much wider AI monetisation strategy for telcos. Developing AI-as-a-Service (AIaaS) offerings will strengthen their value proposition compared to other cloud providers who are heavily reliant on GPU demand for their business model.”

(GUARDIAN)