Network Slicing Moves from Promise to Product, but the Reality Is Narrower than the Hype

The recent announcements of network slicing by operators such as Vodafone, Deutsche Telekom, and Verizon indicate that network slicing is being implemented commercially; however, the actual strategy behind the slicing is far less disruptive. Operators have not strayed from their long-term objective of monetizing a premium product for tightly controlled enterprise and mission-critical organizations in which poor performance or reliability can materially impact their ability to operate, with Service-Level Agreement (SLA)-based options generally enhancing packaging rather than altering the core model.

An example of this is Vodafone's SLA-based slicing service, which is entirely focused on constrained environments such as stadiums, campuses, and transport hubs, where the environment can be engineered and managed to optimize data transmission. Other operators are pursuing similar strategies, including Deutsche Telekom for enterprise communications and telemedicine, Verizon for public safety use cases, and Fastweb for large events requiring telemetry and video data transfer. Each of these examples uses slicing selectively for high-value use cases where failure will have tangible operational or financial impact.

Collectively, it is clear that network slicing is not evolving into a wide-scale monetization mechanism for 5G; rather, it is limited to those use cases where operators can justify the added expense and complexity of guaranteed performance, thereby reinforcing its status as a selective premium layer rather than a widespread enhancement of mobile connectivity.

The recent spate of slicing initiatives appears more to be a recycling of ideas already used for 5G monetization but expressed in a more commercialized way. Telco operators are continuing to pursue many of the same applications for internal enterprise and mission-critical environments with a narrow focus. While SLA-based offers improve the commercial case for slicing, they do not alter the fact that slicing is still a complex, expensive technology that is difficult to scale outside of small, controlled deployments.

Slicing does not eliminate the inherent variability of wireless networks, so operators need to work around rather than eliminate the unpredictability of radio signals from the wireless infrastructure they deploy. The result has been that most slicing deployments still occur in areas where conditions are controlled and managed closely, such as stadiums and large venues like universities, airports, train stations, etc. By deploying their infrastructure in these types of locations, operators can provide reliable performance guarantees without having to make as large a capital investment and incur greater operational complexities.

Adoption has been diverging globally. There is faster growth in Asia-Pacific, where structured, standalone 5G is driving digitalization efforts. In comparison, most enterprises in Western markets do not currently think of dedicated slicing as being required for their existing applications and workloads. Therefore, they continue to prioritize reliability, coverage, and security over slicing-specific capabilities.

Additionally, this is causing slicing to not expand the size of the overall addressable connectivity market. but rather contract it by segmenting this market into a premium tier of a small number of the most used applications. The application of slicing has not yet reached critical mass other than in environments that can guarantee performance. This limits the ability to scale and creates additional reliance on automation and Artificial Intelligence (AI).

Operators need to have another perspective on how they approach network slicing. Network slicing should be viewed through the lens of prerequisites for long-term success: automation, AI-driven orchestration, and real-time network intelligence. Network slicing implementations are currently technically feasible only because the slices are designed and developed to such a high degree and are reliant on having extensive manual processes for configuring and overseeing those network slices. Such implementations may be feasible for a lot of smaller-scale or niche enterprise and public safety customers but cannot support continued growth or become more commercially viable to support the majority of customers.

Currently, operational complexity presents the greatest difficulties in creating slices, not technical feasibility. With no major increase in automation, slicing will continue to be a costly capability that remains in limited use. Telco AI and network automation, along with emerging Agentic AI frameworks can provide significant opportunities. AI-driven orchestrators will enable operators to dynamically create and optimize slices based on network conditions and application requests, all while minimizing manual effort.

To help facilitate a shift to higher levels of performance in delivering slice resources, operators need to establish closer integration between their Service Management and Orchestration (SMO) systems, Radio Access Network (RAN) intelligence, core automation systems, and indoor networks. Operators need to establish autonomous, intentional-based network management systems that allow AI to continually optimize resources to satisfy SLAs on a real-time basis, rather than rely on a static resource configuration process.

Operators also need to rethink the commercial positioning of slice opportunities. Slice capability is most likely to be created in high-value operational environments that directly affect either revenue, safety, or service delivery (such as broadcasting live events, industrial automation, telemedicine, logistics, and public safety communications), rather than through broad-based “premium connectivity” offerings.