Utility metering infrastructure operates under a set of constraints that most IoT verticals don’t share. Devices are deployed in the millions, not thousands. Field life is measured in decades, not years. Physical access to a deployed meter for maintenance or SIM replacement is operationally expensive at the scale of a city-wide or regional rollout. And the regulatory environment governing data accuracy and network reliability varies across jurisdictions, adding compliance requirements on top of the standard engineering ones.
The network architecture decisions made at the start of a smart metering program shape operational outcomes for the entire deployment lifetime. This piece covers the connectivity requirements specific to utility IoT, the technology options and their trade-offs, and the infrastructure considerations that utilities and their technology partners need to get right before deployment locks in.
What Makes Utility Metering Connectivity Different
Most IoT deployments can tolerate a degree of connectivity variability. A fleet tracker that misses a position update is an inconvenience. A smart meter that fails to report a billing reading is a revenue event, a regulatory issue, or both.
Utilities operate metering infrastructure under billing accuracy requirements that translate directly into connectivity reliability specifications. Missed readings require manual estimation, customer dispute processes, and in some jurisdictions, regulatory reporting. At the scale of a large utility with hundreds of thousands or millions of endpoints, even a small percentage of consistently failing devices creates a material operational burden.
The other distinguishing characteristic of utility metering is physical deployment context. Smart meters are installed in locations that optimize for the operational needs of the utility, not for cellular signal quality:
- Underground meter pits for water and gas infrastructure
- Basement electrical panels in multi-unit residential buildings
- Remote agricultural metering points far from dense network coverage
- Industrial facilities with RF-attenuating steel and concrete construction
The cellular technology selected for smart metering must perform reliably in these environments, which means link budget performance is a main selection criterion along with data rate, power consumption, and cost.
Technology Options and Their Trade-offs
NB-IoT is the dominant cellular technology for smart metering deployments globally, and the reasons are well-aligned with the specific requirements of the use case.
NB-IoT’s link budget advantage over LTE-M is approximately 20dB, which translates to meaningful coverage in the challenging physical environments where meters are installed. A signal level that would be insufficient for LTE-M session establishment is often adequate for NB-IoT. In underground meter pit deployments and basement electrical installations, this coverage depth difference is the deciding factor.
Power consumption in PSM at very low transmission frequencies is lower on NB-IoT than LTE-M, extending battery life on battery-backed or battery-only meters. For water and gas meters that have no connection to the electrical grid, multi-year battery life is a hard operational requirement, not a preference.
Per-unit data costs on NB-IoT are lower than LTE-M for the small, infrequent payloads that metering generates. At the scale of a million-endpoint water utility deployment, the per-SIM cost difference between NB-IoT and LTE-M aggregates to a figure that materially affects program economics.
The trade-offs of NB-IoT for utility metering are worth understanding clearly:
- Session establishment latency is higher than LTE-M, which constrains the technology for applications requiring rapid on-demand reads or sub-minute reporting intervals
- Downlink throughput is limited, which affects OTA firmware update delivery timelines for large device populations
- NB-IoT deployment coverage is not yet uniform globally, with some regions still dependent on LTE-M or 4G Cat-1 fallback for areas where NB-IoT is not fully deployed by local carriers
LTE-M is the appropriate choice for smart metering applications where NB-IoT’s constraints create operational problems:
- Advanced metering infrastructure (AMI) deployments requiring on-demand meter reads initiated by the utility head-end system, where NB-IoT’s session establishment latency is too high to meet response time requirements
- Deployments in regions where NB-IoT carrier coverage is incomplete and LTE-M provides the better coverage footprint
- Meters that require regular firmware OTA updates on short timelines, where LTE-M’s higher downlink throughput makes large update packages practical
LoRaWAN is relevant for utility metering deployments where the utility operates its own private network infrastructure and where the extremely low power consumption and low per-device network cost of LoRa are more valuable than the cellular network management that NB-IoT or LTE-M provides. LoRaWAN introduces gateway infrastructure cost and the overhead of managing a private network, which shifts the economics compared to a cellular solution, where the carrier manages the network. For utilities with existing LoRaWAN infrastructure or strong operational technology preferences for private networks, it is a viable metering transport. For utilities without that infrastructure, the operational overhead typically makes cellular the more practical choice.
RF mesh networks using standards like Wi-SUN (IEEE 802.15.4g) are deployed by some utilities as a purpose-built AMI network. Wi-SUN mesh provides high-bandwidth, low-latency connectivity for dense urban meter deployments, with meters acting as mesh nodes to extend coverage. The trade-offs are large capital investment in the mesh infrastructure, geographic constraints on mesh network deployment, and the operational complexity of managing a proprietary mesh network alongside the utility’s other systems. For utilities pursuing greenfield AMI deployments at scale, Wi-SUN mesh and cellular NB-IoT are the two architectures worth serious evaluation.
eUICC and Remote SIM Provisioning for Decade-Scale Deployments
No aspect of smart meter connectivity architecture is more consequential for long-term program success than the SIM strategy, and whether eUICC with remote SIM provisioning capability is incorporated from the start.
The cellular network landscape will change multiple times over the operational life of a smart meter deployed today. Carriers in several markets are already planning or executing 2G and 3G shutdowns. LTE network evolution continues. NB-IoT and LTE-M carrier deployments are expanding in some markets and consolidating in others. A meter with a fixed SIM provisioned for a specific carrier on a specific technology cannot adapt to those changes without physical access to the device, which at utility deployment scale means a truck roll to every affected meter. The cost of that operation at millions of endpoints is prohibitive.
eUICC with remote SIM provisioning via the GSMA SGP.02 specification for M2M deployments, or the more recent SGP.32 for IoT-constrained devices, provides the mechanism to push new carrier profiles OTA without physical device access. When a carrier sunsets a network technology or a utility decides to move its metering traffic to a different carrier for cost or coverage reasons, the transition executes programmatically across the entire deployed estate rather than through a field intervention program.
SGP.32 is relevant for NB-IoT and battery-operated metering devices because it reduces the power overhead of RSP operations compared to SGP.02. On a battery-operated NB-IoT meter in PSM, the power budget available for any single wake cycle is tightly constrained. SGP.32’s IoT-optimized profile management architecture keeps RSP operations within those constraints in a way that SGP.02 does not.
The decision to deploy eUICC-capable meters versus fixed SIM meters is typically a cost decision at procurement time: eUICC-capable hardware carries a premium over fixed SIM equivalents. The calculus changes significantly when the full program lifecycle is included. A single carrier migration executed without eUICC across a one-million-endpoint deployment, at even a conservative truck roll cost per meter, represents a capital exposure that dwarfs the eUICC hardware premium many times over.
Network Architecture for AMI Head-End Integration
Smart meter connectivity doesn’t end at the meter. The data path from meter to utility head-end system involves a connectivity layer, a data management platform, and integration points with billing, outage management, and customer information systems that each carry their own requirements.
The connectivity layer between deployed meters and the utility’s AMI head-end needs to support:
- Bidirectional communication: Most modern AMI deployments require not just meter-to-head-end data push, but head-end-to-meter command delivery for on-demand reads, tamper alerts, remote disconnect and reconnect operations, and firmware update delivery. This bidirectionality requirement affects technology selection, since not all LPWAN technologies support reliable downlink command delivery with the latency profiles that remote disconnect operations require.
- Private APN configuration: Utility meter data transiting public internet segments represents a regulatory and security risk in most jurisdictions. Private APN connectivity, where meter traffic routes directly to the utility’s network infrastructure without exposure to public internet paths, is standard practice for utility IoT deployments of any scale.
- Static IP or device-group addressing: AMI head-end systems typically require deterministic network addressing for meter management operations. Dynamic IP assignment from consumer or standard IoT SIM configurations is incompatible with the addressing requirements of most AMI software platforms.
- Data volume governance: At a million endpoints transmitting multiple times daily, even small payloads aggregate to significant data volumes. Connectivity management platforms that provide pooled data governance, usage monitoring per device group, and automated response to anomalous consumption events are operationally necessary at utility scale.
Connectivity Resilience and Redundancy Requirements
Utility metering occupies a category of IoT infrastructure where connectivity gaps have regulatory implications. Most utility commissions have data accuracy and completeness requirements for billing data, and persistent connectivity failures at a measurable percentage of endpoints trigger regulatory reporting obligations and customer protection processes.
The resilience architecture for smart metering connectivity needs to address two failure modes:
Individual device connectivity failures occur when a specific meter loses cellular connectivity due to network changes, SIM issues, or local RF environment changes. A connectivity management platform with per-device status monitoring and automated alerting surfaces these failures for field investigation before they accumulate into a billing accuracy problem.
Regional or carrier-level network events affect multiple meters at once when a carrier experiences an outage or performs network maintenance. Multi-carrier SIM architecture provides resilience against single-carrier events by enabling affected devices to fail over to an alternate carrier’s network. For a utility with regulatory billing accuracy requirements, this failover capability is the difference between a manageable network event and a reportable data collection failure across a large meter population.
KORE’s connectivity infrastructure supports both resilience requirements through multi-carrier SIM architecture backed by 40+ carrier partnerships across 200+ countries, combined with the ConnectivityPro platform’s per-device monitoring, automated alerting, and network policy controls. With 21M+ active SIMs deployed across utilities and other verticals, the operational experience with large-scale fixed IoT deployments at this level of complexity is reflected in both the infrastructure architecture and the support model.
Planning for the Full Program Lifecycle
Smart metering programs are infrastructure programs, not technology projects. The network architecture decisions made during program design will govern operational outcomes for fifteen to twenty years. The connectivity infrastructure evaluation should be scoped accordingly.
Key decisions that need to be made with the full lifecycle in view:
- eUICC versus fixed SIM, evaluated against the projected cost of carrier migrations and network technology transitions over the deployment lifetime
- NB-IoT versus LTE-M versus hybrid, evaluated against current and projected carrier coverage in the deployment geography
- Private APN and static IP architecture, evaluated against the AMI head-end’s addressing and routing requirements
- Connectivity management platform capabilities, evaluated against the operational scale of managing millions of SIMs across a multi-decade program
The per-unit connectivity cost that looks favorable in year one of a program can look very different when the full lifecycle cost of a fixed SIM architecture is included. Utilities that evaluate connectivity infrastructure as a commodity procurement decision and those that evaluate it as a program-defining infrastructure choice arrive at very different total cost of ownership outcomes.
If your organization is in the planning or procurement phase of a smart metering or utility IoT program, talk to a KORE connectivity engineer about the infrastructure architecture that matches your deployment scale, geography, and program lifetime.
Published June 1, 2026.