How 5G Standalone and Non-Standalone Networks Differ

5G & Telecommunications

•

November 16, 2026

5G networks are often described as either "Standalone" (SA) or "Non-Standalone" (NSA), and that distinction shapes what 5G can actually deliver in a given place. The label on the phone screen might simply say "5G", but the underlying network mode determines latency, reliability, advanced features, and how future-proof the deployment is.

At a high level, Non-Standalone 5G uses 4G LTE as the anchor and adds 5G mainly for faster data, while Standalone 5G runs on a new 5G core built for low latency, network slicing, and large-scale machine-type communication. NSA has enabled a relatively quick rollout of 5G, whereas SA is what unlocks most of the more ambitious 5G capabilities.

The sections below break down how these two approaches differ technically, operationally, and from the standpoint of devices and applications.

What Standalone and Non-Standalone Actually Mean

In 3GPP standards, the terms describe how the radio access network (RAN) and core network are combined.

  • Non-Standalone (NSA) 5G: The 5G New Radio (NR) is added on top of the existing 4G LTE core (Evolved Packet Core, EPC). Control signaling and some user-plane functions still rely on LTE, while 5G is used primarily for high-speed data.
  • Standalone (SA) 5G: Both the radio and the core are 5G. The network uses 5G NR for access and a new 5G Core (5GC) for control and user-plane functions, without depending on LTE.

In practical terms, NSA is a hybrid. The device often maintains a 4G connection as the master link and a 5G link as a secondary data pipe (commonly called EN-DC, E-UTRA NR Dual Connectivity). SA lets a device camp entirely on a 5G cell and use a service-based 5G core that supports more advanced features.

Core Network Architecture: EPC vs 5G Core

Rows of server racks in a data center with visible cables and lights.

The biggest structural difference is the core network.

Non-Standalone: EPC-centric design

In NSA, the existing LTE EPC remains the anchor:

  • The Mobility Management Entity (MME) still handles many control-plane tasks.
  • Serving and Packet Gateways (S-GW and P-GW) manage user-plane traffic.
  • Quality of service (QoS) is based on LTE-era bearers and policies.

A 5G NR base station (gNB) connects to the LTE eNB and EPC. Much of the control logic assumes 4G-style sessions, so introducing new, granular service types is constrained by the EPC’s architecture.

Standalone: service-based 5G Core

In SA, the EPC is replaced with a 5G Core built around loosely coupled, service-based functions. Key shifts include:

  • Split of control and user plane is more refined, with dedicated functions like the Access and Mobility Management Function (AMF) and Session Management Function (SMF).
  • A centralized policy and charging function can slice and manage traffic per user, per device type, or per application more flexibly.
  • Network functions communicate largely via standardized APIs, which supports cloud-native deployment in containers or virtual machines.

This design makes it easier to deploy features like dynamic network slicing, per-application QoS, and fine-grained mobility management that are difficult to retrofit into EPC.

Radio and Spectrum Behavior in SA vs NSA

Both SA and NSA use the same 5G NR air interface, but they use it differently.

NSA radio anchoring

In NSA, the LTE cell usually acts as the primary cell:

  • Control-plane signaling goes over LTE.
  • The device may first attach to LTE, then add 5G NR as an additional bearer.
  • Handover procedures, measurement reporting, and many mobility decisions adhere to LTE-centric logic.

This anchoring can be helpful where LTE coverage is strong and 5G coverage is still patchy, but it also means that some of the latency and overhead of LTE persists even when data flows largely over 5G.

SA radio independence

In SA, the device attaches directly to a 5G cell and uses the 5G core from the start:

  • All signaling and user data can stay in the 5G domain.
  • Spectrum resources can be optimized purely for 5G, including standalone mid-band or millimeter-wave deployments.
  • The network can eventually shrink LTE layers or refarm LTE spectrum for NR without depending on 4G anchoring.

SA is particularly important for bands deployed exclusively for 5G, where building a future network without long-term LTE dependence is part of the strategy.

Latency, Reliability, and Throughput Differences

Both SA and NSA can offer high peak speeds, but their behavior under load, in mobility, and for time-sensitive applications diverges.

Latency

NSA typically exhibits lower latency than LTE-only networks thanks to 5G NR’s more efficient frame structure, but the dependence on EPC and LTE signaling adds extra steps. Session setup, mobility signaling, and control-plane procedures may still run through LTE/MME paths.

SA can reduce end-to-end latency further because:

  • Control and user plane are optimized in the 5G Core.
  • The core can be distributed closer to the edge, reducing the physical path length.
  • The network can create dedicated low-latency slices with specific resource reservations.

For most ordinary smartphone activities, both modes feel similar once latency is already in the tens of milliseconds range. For industrial control, real-time gaming at scale, or remote operation, the tighter bounds of SA matter more.

Reliability and consistency

NSA can benefit from dual connectivity: if the 5G link fades, the LTE anchor still carries data. This can smooth handovers and provide graceful degradation in mixed-coverage areas.

SA depends solely on 5G coverage. Where that coverage is dense and well-engineered, it can deliver very consistent performance, especially when combined with slicing and prioritized QoS. In sparse or early-stage deployments, coverage gaps may be more noticeable unless multi-layer 5G deployment is in place.

Throughput

Peak data rates are often similar in both modes when the same spectrum and radio features are used. Differences tend to arise from:

  • How much spectrum is available for 5G versus LTE.
  • Scheduler behavior in dual connectivity (NSA) compared with pure NR scheduling (SA).
  • Backhaul and core network capacity planning.

In practice, NSA commonly delivers a noticeable jump in throughput compared with LTE, while SA’s main gains are in latency, flexibility, and consistency under diverse traffic patterns rather than headline peak speeds alone.

Network Slicing and Advanced 5G Services

One of the headline ideas in 5G design is network slicing: carving the physical network into multiple virtual networks with different characteristics and policies. This is where SA and NSA differ sharply.

NSA limitations for slicing

Some slice-like behaviors can be emulated in NSA using QoS differentiation and separate APNs (Access Point Names), but the capabilities are constrained by EPC’s bearer-based design. Fine-grained, dynamic, end-to-end slices (from device through RAN, transport, and core) are difficult to implement robustly.

SA as the foundation for full slicing

SA is the practical baseline for full network slicing because the 5G Core:

  • Represents slices as first-class entities with their own control and user-plane instances.
  • Can route traffic differently per slice, including to specific edge computing sites or application servers.
  • Supports lifecycle management of slices: creation, scaling, modification, and deletion, often in an automated way.

This enables scenarios such as:

  • A slice for consumer mobile broadband with best-effort performance.
  • A slice for emergency services with strong prioritization and resilience.
  • A slice for factory automation with stringent latency and jitter requirements.

Similarly, ultra-reliable low-latency communication (URLLC) and massive machine-type communications (mMTC) are much more naturally supported on SA, where QoS control, device density management, and scheduling rules can be fully aligned with 5G Core capabilities.

Deployment Strategies and Migration Paths for Operators

From an operator’s perspective, NSA and SA are stages in a multi-year evolution rather than simple alternatives.

Why NSA was often deployed first

Many operators started with NSA because:

  • It reuses the existing EPC and LTE sites, reducing upfront complexity.
  • It gets a 5G icon on devices quickly, meeting market expectations.
  • It allows incremental rollout of 5G radios without immediately overhauling the core.

This approach lets operators gain early 5G experience, tune radio planning for new bands, and build backhaul capacity while deferring large-scale core transformation.

Transitioning to SA

Moving to SA involves several steps that typically include:

  1. Deploying a 5G Core in a cloud-native environment, often on virtualized or container-based platforms.
  2. Integrating the 5G Core with existing OSS/BSS (operations and business support systems) and policy engines.
  3. Enabling dual-mode RAN that can serve both NSA and SA users on the same radio hardware, where supported.
  4. Gradually onboarding enterprise and advanced services to SA slices while consumer traffic may still be mixed.

The hybrid stage, where NSA and SA coexist, can last for years. Traffic steering policies decide which subscribers or use cases attach through SA versus NSA, depending on device support, coverage, and service requirements.

Device Support and User Experience

Devices are another key differentiator between SA and NSA in real-world usage.

NSA compatibility

Most early 5G smartphones were built with NSA as the default mode because operators initially offered mainly NSA networks. On such devices:

  • The modem maintains LTE and NR connections concurrently.
  • Voice typically runs on VoLTE (LTE-based voice over IP), even when data uses 5G.
  • Battery drain can increase if dual connectivity is not efficiently managed, since the device keeps multiple radios active.

SA capabilities

Devices that support SA can register directly with a 5G Core and may support features such as:

  • Voice over New Radio (VoNR), where voice calls remain entirely on 5G.
  • Enhanced power saving through 5G-specific mechanisms like more flexible DRX (Discontinuous Reception) tuned for SA behavior.
  • Access to dedicated slices for specific enterprise or premium services.

On the user side, the differences can be subtle initially. Web browsing or video streaming might not feel dramatically different. As more SA-only features like specialized slices or URLLC-based applications become available, devices without SA support may be limited to basic connectivity on NSA or fall back to LTE in areas where operators favor SA.

Security and Network Management Differences

Security and operational management also evolve between NSA and SA.

Security in NSA

In NSA, many security mechanisms remain rooted in LTE designs:

  • Authentication and key management are anchored in the 4G core.
  • Some signaling protection relies on LTE-era protocols and cipher suites.
  • The mixed architecture can create more complex trust boundaries, as both LTE and NR need to be secured and coordinated.

While robust, this structure was not originally designed with large-scale virtualization, slicing, and continuous software lifecycle management in mind.

Security model in SA

SA’s 5G Core is built with more explicit provisions for:

  • Unified authentication and authorization across diverse device types, including IoT and vertical-specific equipment.
  • Better isolation between slices, so compromising one slice does not automatically expose others, assuming correct configuration.
  • Modern security practices aligned with cloud-native deployments, such as more granular access control between microservices and improved observability.

From a management perspective, SA networks often rely heavily on automation and analytics for functions like scaling, healing, and optimization. NSA can benefit from these tools as well, but SA’s architecture is aligned with them from the outset.

Enterprise and Private Network Use Cases

A factory floor with large machines and visible 5G equipment.

Enterprises considering private or campus networks face a different set of choices when comparing SA and NSA.

NSA in enterprise settings

Using NSA for private networks is possible but typically involves some dependence on an external LTE core or integration with a public operator’s infrastructure. This may be adequate for:

  • Enhanced mobile broadband across a campus where LTE already exists.
  • Transitional environments where legacy LTE equipment must be preserved.

However, the tight integration with LTE and EPC can limit how deeply the enterprise can customize QoS, slicing-like behavior, or on-premises processing.

SA for tailored enterprise networks

SA lends itself more naturally to:

  • Fully on-premises 5G cores that keep sensitive data local.
  • Dedicated slices for safety-critical applications, such as real-time machine control or augmented reality support in manufacturing.
  • Integration with edge computing platforms hosted within the enterprise site.

Because SA allows all signaling and user traffic to remain within the 5G domain, enterprises gain more flexibility in traffic engineering, security policies, and service differentiation compared with NSA-based designs.

How Networks Are Likely to Evolve

The long-term direction of 5G deployment influences the relevance of SA and NSA over time.

Several trends can already be seen:

  • As 5G coverage matures, operators tend to introduce SA first in urban or high-demand areas where low-latency and enterprise services are viable, while keeping NSA in wider coverage zones.
  • New spectrum bands are often planned with SA in mind, particularly where there is no existing LTE layer to anchor NSA.
  • Over time, as more traffic and services migrate to SA and VoNR becomes widespread, the rationale for maintaining extensive LTE coverage and EPC infrastructure decreases.

In parallel, standards continue to evolve. Later 3GPP releases add enhancements that are easier to implement on SA, including refined slicing features, better support for time-sensitive networking, and deeper integration with non-terrestrial networks such as satellites. NSA may continue as a compatibility layer for some years, but most advanced 5G capabilities are expected to center on SA deployments.

Making Sense of SA vs NSA for Planning and Expectations

The distinction between Standalone and Non-Standalone 5G is not just a technical curiosity. It affects deployment timelines, what applications are realistic, and how networks will be upgraded.

NSA has been the practical stepping stone: a way to deliver higher speeds by bolting 5G radios onto the existing 4G core. SA is the more comprehensive transformation: a new, service-based core that supports slicing, low-latency services, and more flexible integration with edge computing and enterprise networks.

For network planners, recognizing how each mode shapes spectrum decisions, device requirements, and service roadmaps is central to long-term strategy. For users and organizations relying on connectivity, understanding whether a given service runs over NSA or SA helps set reasonable expectations about performance, capabilities, and how readily future 5G features can be adopted.

Knowledge Tags 🏷

5g networkstelecommunicationsnetwork architecture

Frequently Asked Questions

Find quick answers to common questions about this topic

Many operators run both modes in parallel during a transition phase. The same physical 5G radio sites can often support NSA and SA simultaneously, with different core networks handling each mode. Devices attach using the mode they support and that the operator has configured for their subscription, which allows gradual migration without a disruptive cutover.

Standalone 5G does not automatically guarantee higher peak speeds than Non-Standalone. Throughput depends heavily on how much spectrum is used, which bands are deployed, and how the radio layer is engineered. SA’s main advantages are lower latency, better consistency, and access to advanced features like slicing rather than raw data rate alone.

LTE is expected to remain in service for many years as a coverage layer and compatibility anchor, even where 5G Standalone is available. Over time, some spectrum now used for LTE may be refarmed to 5G NR as more devices support SA and VoNR. The speed of this shift depends on local device penetration, regulatory conditions, and each operator’s strategy.

Most phones only display a simple 5G icon and do not explicitly label SA or NSA on the main screen. Some devices provide this detail in engineering menus or diagnostic apps that expose connection parameters like NR-NSA or NR-SA mode. Operators occasionally publicize where SA is available, but in day-to-day use the distinction is usually invisible unless a specific SA-only feature is being used.

Not all 5G-capable IoT devices support Standalone operation. Some modules were designed for early NSA deployments and rely on LTE anchoring. Newer IoT chipsets more often include SA support, which is important for devices that need advanced features, tight latency control, or connection to private 5G cores. Checking the supported 3GPP releases and modes is necessary when planning large-scale IoT deployments.

Private 5G networks do not always require Standalone mode, but SA is usually favored for fully controlled on-premises deployments. NSA-based private networks tend to rely on integration with an external LTE core or public network, which can be suitable for less demanding use cases. For tightly controlled latency, security, and local breakout of traffic, SA is typically the more suitable architecture.

Expectations for future generations suggest a move toward more unified, cloud-native cores rather than long-term reliance on older technologies. Early phases of new generations sometimes use non-standalone arrangements for compatibility, but as capabilities mature, standalone architectures usually become central. The pattern seen with 5G hints that any 6G systems would eventually focus on standalone-like designs once deployment is widespread.

About the author

Victor Okafor

Victor Okafor

Contributor

Victor Okafor is a visionary AI ethics specialist with 14 years of experience developing responsible implementation frameworks, algorithmic accountability systems, and governance structures for artificial intelligence applications across diverse sectors. Victor has helped numerous organizations integrate AI ethically through his practical evaluation methodologies and created several widely-adopted approaches to balancing innovation with responsible deployment. He's passionate about ensuring technology serves humanity's best interests and believes that ethical considerations must be built into AI systems from inception rather than added afterward. Victor's thoughtful perspective guides developers, business leaders, and regulatory bodies working to maximize AI's benefits while minimizing potential harms.

View articles