As the telecommunications industry evolves, the migration of telco workloads to public cloud platforms presents both opportunities and challenges.
The STL Partners study, Hyperscalers in the telco vertical: The future is hybrid, highlights that while telcos have progressively moved IT workloads, such as business support systems (BSS), to the cloud over the past decade, a more cautious approach persists regarding network workloads. This hesitance is particularly evident in the migration of critical components like the 5G core, IMS, and radio access network (RAN) systems.

“The migration of 5G core, IMS and RAN workloads to the public cloud remains nascent,” says George Glanville, senior analyst at STL Partners.
The study reveals that there have only been 12 production deployments of 5G core workloads on public cloud platforms, mainly among greenfield networks.
For RAN deployments, the situation is even more limited, with just three known instances, one of which has since been decommissioned. This limited adoption underscores the significant hurdles telcos face in transitioning to cloud-based network solutions.
Several factors contribute to this reluctance. Key concerns include the technical and commercial performance of cloud models and a misalignment with telco investment cycles.
Glanville points to uncertainties surrounding the commitment of virtual network function (VNF) and cloud-native network function (CNF) vendors to support these deployments. Moreover, governmental pushes for digital autonomy discourage reliance on non-sovereign infrastructure, complicating the decision to migrate network workloads.
In contrast, the area of AI workloads has seen greater success. STL Partners reports 88 instances of telco-hyperscaler AI partnerships, with hyperscalers becoming vital allies in telco AI initiatives. The research also identifies 20 collaborations focused on sovereign cloud services, primarily led by Huawei Cloud.
“Hyperscalers will form one of the several pillars supporting the hybrid cloud on which networks will operate,” explains Glanville.
While BSS workloads may predominantly utilise public cloud for its performance and cost benefits, more demanding network tasks like RAN will likely remain on private telco clouds for the foreseeable future.
The extent of network function deployment on hyperscale platforms will depend heavily on the cloud providers’ commitment to these relationships.
As Glanville warns, “There is substantial investment required on hyperscalers’ part to host demanding network workloads.”
He cites Microsoft Azure’s recent disbanding of Azure for Operators as a cautionary tale, illustrating the potential for hyperscalers to redirect capital away from telco verticals amidst the ongoing AI boom.
In summary, while the telco industry is making strides in cloud migration, significant challenges remain, particularly regarding network workloads. The success of future collaborations between telcos and hyperscalers will hinge on clear commitments and sustained investment.