What Is 5G Network Slicing? Architecture, Use Cases, and Monetization

5G network slicing is a technique that operators use to divide a single physical 5G network into multiple separate virtual networks, called slices—each configured for a different job. The operator’s charging systems and business software price and bill what runs on those slices—per user, device, session, usage event, or a bundled package, depending on the deal.  

That gives an operator a range of services it can sell from one network, each configured for a specific performance need: a low-latency slice for a factory floor, a high-reliability slice for a hospital’s monitors, a high-uplink slice for a stadium’s live broadcast, or a massive-scale slice for a utility’s smart meters.  

This guide covers what slicing is, how it works, common 5G slice types, the architecture behind it, real-life deployments, and how operators turn a slice into revenue. 

What is 5G network slicing? 

5G network slicing is the practice of running many separate virtual networks over one physical 5G network, each configured for the bandwidth, latency, and reliability a particular business or service needs. Every slice shares the same infrastructure—tower, fiber, and 5G core—but stays logically isolated from the others so the operator can track what happens on each one and charge for it.   

Picture the 5G physical network as one highway divided into lanes—an express lane, a freight lane, and a local lane. In slicing, those lanes are the slices, each set up for a specific purpose. But the lanes aren’t permanent: they’re created with software, so an operator can widen one, narrow another, or open a new lane for a few hours and close it again. And just as you don’t build a separate highway for every type of traffic, an operator using slicing can add a service without building a new network. 

Adding a slice costs far less than building a network, though it isn’t free. A slice can require extra capacity, software licenses, orchestration, assurance tooling, testing, integration, security controls, and the operational support to run it. What an operator gets for that is the ability to sell a service it couldn’t sell before, and when a slice carries a performance guarantee, it can also charge a premium for it. That’s the business of slicing: one network, a range of services, priced for what each one delivers.  

How does 5G network slicing work?  

5G network slicing works by using software to coordinate network resources for each slice across the radio, transport, and 5G core. Depending on how the operator sets it up, a slice might get dedicated capacity, capacity shared with other customers, or priority when several slices compete for the same capacity.  

How software makes slicing possible 

Slicing relies on several technologies working together. Software-defined networking (SDN) and network functions virtualization (NFV) are two important enablers.    

• SDN gives operators programmable control over how traffic is steered through parts of the network. This makes it easier to apply different routing and resource policies to different services without configuring every network device individually. 

NFV allows network functions such as mobility management, policy control and session management to run as software instead of requiring a dedicated hardware appliance for each function. This gives operators more flexibility to deploy and scale those functions as demand changes. 

Operators already used SDN and NFV in 4G networks. What 5G adds is a standards-based core architecture designed to identify, select and manage network slices as part of the service delivered to a device or customer.  

How a device lands on the right slice 

A slice is identified for selection purposes using Single Network Slice Selection Assistance Information, or S-NSSAI. An S-NSSAI contains a Slice/Service Type and may also include a Slice Differentiator. When a device registers, the 5G core evaluates the slices requested by the device against subscription information, network availability and operator policy. The AMF, with support from the Network Slice Selection Function where required, determines which slices the device can access. 

From there, the slice has to hold up across every part it travels. 

Slicing runs end-to-end 

A lane only helps if it's clear the whole way. The slice has to be honored across the entire network, or it can't deliver the performance the customer is paying for. The service may cross the radio access network (RAN), transport network and 5G core and, in some cases, edge or cloud infrastructure. 

Each part handles the slice separately, so any one of them can jeopardize the slice’s performance. If traffic gets held up crossing the transport network, for example, that time is lost and the core can’t make it up. 

If one part doesn’t apply the slice’s treatment, the traffic still gets through—it just moves alongside ordinary unsliced traffic, without the quality of service (QoS) the customer is paying for. That's why a slice is only as strong as its weakest part: skip one, and the slice can’t deliver the promised treatment no matter how well the rest performs.  

5G network slicing requires 5G standalone (SA) 

Full end-to-end network slicing needs a 5G standalone (SA) architecture: 5G radio connected to a purpose-built 5G core. The 5G core provides the native slice-selection, session-management, policy and charging capabilities required for standards-based slicing. The RAN, transport and management systems must also support the service requirements. Without a 5G core, an operator cannot use the complete native 5G slicing framework, although it can still apply other forms of traffic differentiation and prioritization. 

Non-standalone (NSA) 5G runs 5G radio on the older 4G core, which was never built to slice traffic end to end—so operators get 5G speed, but not the end-to-end isolation and control a slice requires. Standalone 5G swaps in the native 5G core, and only then can a slice be identified by its own S-NSSAI and honored across all three parts of the network. 

For the full NSA-to-SA migration path, see the 5G deployment guide. 

BY THE NUMBERS: According to the June 2026 Ericsson Mobility Report, some 390 service providers have launched 5G, but fewer than a quarter have launched standalone 5G. Without that standalone core, operators can’t offer full end-to-end 5G network slicing with per-slice charging and SLA-backed performance. 

How 5G network slicing architecture sets up monetization  

Getting a slice across the network is only part of what makes it work. Sitting above the network layer are the systems that manage and sell slices. From a commercial perspective, it is useful to think about slicing in three broad layers: the monetization layer that turns slices into revenue, the automation layer that manages them, and the network layer—the RAN, transport network, and 5G core—at the base.  

The monetization layer is where a slice becomes a product. It contains the business support systems (BSS) that define the commercial side of the deal: what the operator promises in a service-level agreement (SLA), what it costs, and how it’s charged and billed. The BSS also handles taxation, revenue assurance, and financial reporting. Determining whether the SLA was delivered happens in the automation layer, which comes next.  

The automation layer is where slices are built and managed, through orchestration and assurance. It builds a slice across the radio, transport, and core, then continuously measures whether the slice is delivering what the SLA promised. When performance slips, it can take action to bring the slice back. That measurement is also what tells the operator whether the commitment was met. Without the automation layer, an operator could only run a few slices, each one hand-configured. With it, the same operator can run many at once.  

The network layer is the physical and virtual infrastructure a slice runs on: the RAN, the transport network, and the 5G core. All three have to apply the slice’s treatment, or the operator’s guarantee fails.

That architecture lets an operator build a slice for almost any need. In practice, slices are standardized into three types—each built for a different kind of service. 

Common 5G network slice types: eMBB, URLLC, and mMTC 

Not every service wants the same thing from a network. A video stream wants bandwidth; a factory robot wants split-second response; a field of sensors just wants to check in cheaply, by the thousands. So 5G slicing is built around three common slice types, each configured for one of those priorities—bandwidth, latency, or scale. All three come from 3GPP, the body that sets global mobile standards, in its Release 15 specifications. 

Enhanced mobile broadband (eMBB) prioritizes bandwidth 

eMBB slices are the high-bandwidth type. A common case is a dense, crowded place like a stadium, where thousands of people and devices hit the network at once. In practice, most slices sold commercially today are eMBB variants—a premium or prioritized broadband tier rather than one of the more specialized types. 

Ultra-reliable low-latency communications (URLLC) focuses on latency 

URLLC is designed for stringent latency and reliability requirements, where a delayed or dropped signal can trip a safety system or stop a production line. Commercial uses today include industrial control, grid protection, ports, and automated guided vehicles (self-driving carts that move materials around a warehouse or port). Self-driving cars and remote surgery get cited more often, but they’re mostly still in pilots. 

Massive machine-type communications (mMTC) manages scale 

mMTC is the scale type: tens of thousands of low-power devices, each sending a small reading now and then, on batteries designed to last for years.  

The key benefits of 5G network slicing  

Everything this article has covered so far—what slicing is, how it works, and the architecture—leads to a practical question: What does an operator gain?  

The main benefits of 5G slicing are: 

• New revenue streams: An operator can charge more for a customized slice with a guarantee than for a flat data plan, because that guarantee can command a premium. 

• Enterprise customization: Instead of the same connection for every customer, each customer can buy a slice tailored to the exact performance they need—no more, no less. 

• Faster time-to-market: A new service can launch in days instead of the many months it takes to build dedicated physical infrastructure. 

• Resource efficiency: An operator can run many revenue streams from one network instead of building and running separate networks for each.  

• Reduced operational costs: Since slices are set up and managed in software, running them takes less manual, on-site work than standing up separate networks. 

• Improved quality of service (QoS): A slice can be apply for the differentiated QoS, priority and resource policies a service needs, and keep delivering it even when another customer’s traffic surges.  

5G network slicing use cases and real-world examples

Network slicing is used wherever best-effort connectivity isn't good enough—factory automation, live broadcasting, connected healthcare, asset tracking in logistics. And it's moved from the lab into products operators are selling today. The four commercial deployments below show the range. 

 These aren't the only launches, but adoption is uneven across industries. Broadcasting and live events have real commercial slices today—KDDI in Japan sells a broadcast slice used by TV crews at stadiums, and UK operator BT has run multiple concurrent slices at live sporting events. Manufacturing is close behind, mostly through standardized "campus" slice plans like Vodafone's. But in healthcare and logistics, almost everything is still a pilot: hospitals and ports are testing slicing in trials and proofs-of-concept, not necessarily buying it as a standing service. The technology works across all four industries; what varies is how far each industry has moved from trial to product. 

Network slice orchestration: Managing the slice lifecycle 

Orchestration is the software that manages each slice through its whole life—design, launch, adjust, retire. It’s what lets an operator run hundreds of slices instead of configuring each one by hand.  

The four stages of a 5G slice’s life (according to 3GPP) 

• Preparation: The operator designs the slice and its template, then confirms the network can support it before anything is built. 

• Commissioning: The slice is created, and its resources are allocated across the RAN, transport, and core.  

• Operation: The slice goes live—monitored and adjusted as demand shifts. This is also the stage where the slice gets charged and generates revenue.  

• Decommissioning: The slice is shut down, and the capacity it was holding is freed up for other slices to use. 

From slice requirements to a deployed slice: GST, NEST, NSI and NSSI 

The GSMA Generic Network Slice Template (GST) defines a common set of attributes that can be used to describe network slice requirements. A Network Slice Type, or NEST, applies values to those attributes for a particular service or customer requirement. 

A Network Slice Instance (NSI) is the deployed logical network that provides the required capabilities and characteristics. A Network Slice Subnet Instance (NSSI) represents a constituent part of that slice, such as a RAN, core or transport subnet, and may be shared across more than one NSI depending on the design. 

 Orchestration translates service requirements into the network functions, subnets, policies and resources needed to create and operate the NSI. Assurance then monitors whether the service continues to meet its defined objectives. 

Static vs. dynamic network slicing 

Because slices are software-defined, an operator can change what a slice gets. A static slice keeps its configuration until someone adjusts it. A dynamic slice can be modified, scaled or reconfigured automatically or on demand as service requirements change. 

A static slice is one an operator designs, provisions using orchestration, and leaves running—such as a standing slice for a hospital or a factory that keeps the same configuration day to day. It's simple and predictable, but it can't adapt, so one slice can sit half-idle while another is overloaded—unless an operator changes it. Many early commercial offers use relatively fixed configurations, while more dynamic and on-demand models remain less mature. 

Dynamic slicing reallocates capacity in real time based on demand—a dynamic slice can expand for a live event and shrink afterward, or the network can rebalance between slices as traffic moves through the day. It's more efficient, but harder to run—it takes advanced software to move capacity between slices automatically, and few networks can do that at scale yet.

Both static and dynamic slices are still set up by the operator. Exposed slicing hands that control over to the customer—covered under monetization, where slicing becomes a self-service product. 

Challenges of 5G network slicing implementation 

Slicing asks an operator to run new software, split limited capacity, and manage multiple vendors—all before the first slice pays for itself. 

Operational complexity 

Running many end-to-end slices in software—work that used to live in dedicated hardware—takes orchestration and management tools that many operators haven't used before. 

Resource contention 

Slices share one physical network, so a fixed pool of capacity has to be divided carefully. Give too much to one slice, and another suffers; give too little and it can’t deliver what the customer was promised. Matching what each slice gets to what it actually needs is ongoing work, not something an operator sets once and leaves. 

Multi-vendor interoperability 

A slice crosses RAN, transport, and core equipment that often comes from different vendors, and getting it to behave identically across all of them is hard. The RAN can be particularly challenging because radio resources change continuously with coverage, interference, mobility and demand, while vendor implementations and management interfaces may differ.  

An operator running Ericsson in one region and Nokia in another is relying on two different pieces of software to deliver the same service consistently—and the industry hasn't fully standardized how to make them match.  

High upfront cost 

Implementing a 5G core, orchestration systems, and updated business support systems carries a hefty upfront cost—and that’s before the slice-based revenue arrives to justify it. 

Provisioning risk 

A slice can fail quietly. It’s still connected and passing traffic, but no longer getting the treatment needed to meet the guarantee the customer paid for. Its traffic may simply move alongside ordinary unsliced traffic instead. Because the service may remain connected while performance degrades, reliable provisioning, telemetry and continuous assurance are critical to detecting problems before they become SLA breaches. 

Security demands 

Logical isolation can help limit the impact of some security incidents, although shared infrastructure, network functions and management systems still require appropriate security controls. A slice that's assumed to be secure, but not checked, is a risk, not a guarantee. 

None of these is a reason to avoid slicing—they're the reasons to approach it as a serious operational and business commitment rather than a feature to switch on. 

5G network slicing monetization: From technology to revenue 

A slice that runs but can't be billed is a cost, not a product. That's the part most explanations of slicing skip: the network makes the slice possible, but the operator's business systems—its BSS—turn it into something a customer can pay for. Operators define what the SLA promises, track what the slice uses, and produce the bill.  

What a slice earns comes down to a few decisions: who the operator sells it to, how it charges for it, and whether its billing systems can keep up. Three business models: B2B, B2C, and B2B2X provides the underlying network capability, but the commercial service may be sold directly by the operator or through a partner, and that shapes how the slice gets packaged, priced, and billed. There are three arrangements: 

• Selling to a business (B2B): The operator sells a slice straight to an enterprise—a factory, a hospital, a broadcaster—that wants guaranteed connectivity for its own operation. 

• Selling to a consumer (B2C): The operator folds a slice into a consumer plan, like a premium tier for gaming or streaming, so the customer gets better performance without ever hearing the word "slice." 

• Selling through a partner (B2B2X): The operator sells a slice as a building block to another company—a cloud-gaming platform, a connected-vehicle service—that packages it for its own customers. This is where operators see growth potential, because it reaches markets it couldn't serve on its own. It's also where billing gets hardest, since the revenue has to be measured and split between the operator and the partner. 

Three ways to charge for a slice 

Once the buyer is settled, the next question is how to bill them. The charging model depends on what is being sold, who is buying it and how usage, access, performance, capacity or duration is measured. 

• One slice, billed on its own: The simplest case—charge for a single slice by the devices on it, the data it uses, or specific events. A utility might pay a flat per-device rate for its meters' everyday reporting, plus a one-time charge the day it needs a burst of extra capacity. 

• Several slices, one bill: A single enterprise often runs more than one slice—perhaps one for the factory floor and another for its office network—each configured for a different job. The operator puts them on one account and bills them together. 

• Through a partner: In the reseller case, the operator doesn't bill the end customer at all. It bills the partner—a cloud-gaming platform, perhaps—and the partner bills the people actually using the service. 

Slice-based services can support a wider range of commercial models, including fixed subscriptions, usage-based charging, per-device pricing, event charges, capacity reservations and SLA-linked pricing. 

Exposed slicing and network as a service (NaaS) 

The most advanced model lets the customer buy and run a slice themselves. Exposed slicing opens slice controls to enterprises and developers through network APIs—software interfaces that let a customer's own systems request a slice directly, with no sales call and no manual setup. This is what the industry calls network as a service, or NaaS: buying connectivity on demand the way a company buys cloud computing, instead of negotiating a contract for it. 

It turns slicing from something the operator configures into something customers order like any other product. It's the least mature model today, but it's the clearest picture of where slice monetization is heading. 

Billing has to be ready when the network is 

Slices sold with a performance commitment behind them are SLA-backed products, sold across three different business models. Pricing can vary by slice, usage, device, event, or commercial arrangement. Billing for them can involve adding new capabilities to the systems an operator already runs—new ways to price, meter, and settle revenue with partners. That work has to happen while the network is being built, so the operator can sell a slice as soon as the network can deliver one.  

Where the market stands 

Commercial slicing is past the demo stage and into real services—the number of commercial slicing-based connectivity offerings climbed from 65 in Ericsson's November 2025 Mobility Report to 84 in the June 2026 edition.  

ABI Research projects the slicing market growing from about $6.1 billion in 2025 to $67.5 billion by 2030, with enterprise being the larger share. GSMA Intelligence finds that services beyond core connectivity are still adding only about two percentage points of revenue growth for major operators—the new revenue that slicing enables is small so far.  

How CSG helps operators monetize 5G network slicing  

A slice only becomes a product when the business systems can price, meter, and bill it—and that commercial layer is what CSG builds.  

CSG’s 5G charging enablement suite (CHF, CGF, IWF, PCF) supports converged charging for network slicing, B2B2X, IoT, subscription and edge-based services, and multi-access edge computing, so an operator can set SLA-based, usage-based, or capacity-based pricing, and build revenue models beyond traditional subscriber billing.  

For operators running legacy and 5G environments, CSG Convergent Mediation & Activation bridges charging flows between the two. It routes and transforms usage and charging data between network systems and the appropriate downstream charging or billing platforms, allowing legacy and 5G environments to coexist. It allows an operator modernize in phases instead of rip and replace.  

For operators seeking a cloud-native SaaS monetization platform, CSG Ascendon supports subscription, usage-based, hybrid and multi-party billing on a single platform, and can launch new digital services in weeks. It can serve as the billing and monetization layer for new slice-based and B2B2X offers, alongside network-side charging, policy, and mediation.

Frequently asked questions 

What is the difference between network slicing and a VLAN? 

A VLAN (virtual LAN) splits one local network into separate lanes of traffic—useful, and it can prioritize traffic, too, but only inside that one network. A network slice runs across the whole mobile network, from the RAN through the core, and can carry a service-level agreement covering bandwidth, latency, and reliability. So, a VLAN keeps traffic separate inside one network; a slice delivers a service across the entire mobile network, with a commitment behind it when the customer is paying for one. 

What is 5G SA network slicing? 

True end-to-end slicing requires 5G standalone (SA)—5G radio connected to a native 5G core. Non-standalone (NSA) 5G runs 5G radio on a 4G core that was never built to slice traffic end-to-end, so it delivers 5G speed but not the isolation or service treatments a slice needs. 5G SA network slicing is slicing done on that standalone architecture, where a slice can be honored across all three parts of the network—the RAN, transport, and core. 

Why is network slicing important? 

Because it lets an operator sell more than plain connectivity. Instead of giving every customer the same best-effort network, the operator can offer a slice built for exactly what one customer needs—and can put a guarantee behind it. That's something a customer will pay more for, and it reaches businesses a standard data plan can't serve: a factory that needs reliability, a broadcaster that needs an uplink, or a cloud-gaming platform that needs low latency. 

What are the challenges of 5G network slicing? 

Running many slices at once takes automation that many operators don't have yet. Each slice has to meet its SLA end to end—across the RAN, transport, and core, often stitched together from several vendors' equipment. And the billing side has its own challenges: charging per slice against an SLA, and settling revenue with partners, can mean extending many of the systems an operator already runs.  

What is the difference between static and dynamic network slicing? 

A static slice uses configurations and resource policies that remain relatively stable during operation. A dynamic slice can be scaled or reconfigured automatically or on demand as requirements change. Static configurations are generally simpler to operate, while dynamic slicing requires more mature orchestration, assurance and policy automation. 

How do operators monetize 5G network slicing? 

Monetization runs through the operator's billing systems—its BSS (business support systems)—which price, meter, and bill each slice. Operators sell slices under three models (B2B, B2C, and B2B2X) and charge in ways that fit how each slice is used—metering by device or by event on a single slice, or billing a partner who resells to the end customer. Pricing can be based on subscription, user, device, usage, event, duration, capacity, service level or partner arrangement, depending on the offer.