Enterprise and Industrial IoT Connectivity Solutions: Five Beliefs That Do Not Survive Deployment

Most enterprise IoT connectivity projects begin with a radio technology question: NB-IoT vs. LTE-M vs. LoRaWAN vs. satellite. Engineers compare power, throughput and coverage, then record a decision. That work matters, but it is often the least consequential decision in the program.

The radio is relatively reversible. A different bearer can often be introduced during a hardware refresh. What is much harder to change is the SIM soldered into thousands of devices, the provisioning model, the IP architecture built into customer firewall rules, and a long-term connectivity contract signed before the real data profile is known.

IoT Analytics reported that global cellular IoT connections grew 13.3% in 2025 to 4.7 billion, the slowest growth rate it has recorded since 2020. The same research counts roughly 21.1 billion connected IoT devices overall. As the installed base grows and expansion slows, IoT increasingly becomes a migration problem. IoT Analytics cellular IoT market update

The following five beliefs show where enterprise assumptions tend to break down after deployment.

Belief One: The Connectivity Decision Is Mainly a Technology Decision

The technology choice is important, but the device and deployment pattern usually narrow the options quickly.

Cellular 4G and 5G

Used for high-throughput or low-latency applications such as video telematics, connected cameras, gateways and mobile routers. These devices are generally mains- or vehicle-powered.

LTE-M

LTE-M, or Cat-M1, supports moderate throughput, lower power consumption and mobility between cells. It is commonly suited to moving assets.

NB-IoT

Designed for stationary sensors sending small amounts of data, including devices installed in basements, cabinets and other difficult locations. Its mobility support is more limited than LTE-M.

Unlicensed LPWAN

Technologies such as LoRaWAN operate on unlicensed spectrum and use gateways that the operator owns. This can work well across controlled sites such as plants, campuses and utility areas.

Satellite and Non-Terrestrial Networks

Satellite connectivity serves assets outside reliable terrestrial coverage, including maritime, mining, pipelines and remote agriculture.

The real issue is availability. A technology can be technically ideal but commercially unavailable where the device will operate. The GSMA lists 129 commercial LTE-M networks and 140 commercial NB-IoT networks worldwide, but availability varies significantly by country. GSMA Mobile IoT commercial launches

Belief Two: The Coverage Map Tells You Where Devices Will Work

Coverage maps show radio propagation, not whether an enterprise can commercially use that network for the lifetime of its devices.

Four issues sit between a green coverage area and a working industrial deployment:

• Roaming permission: Global connectivity providers often depend on roaming partners whose agreements can change.
• Permanent roaming: Some countries restrict or prohibit devices from permanently roaming on foreign profiles.
• Technology-specific roaming: NB-IoT roaming remains less mature than LTE-M roaming in many regions.
• Network retirement: 2G and 3G shutdowns have removed fallback networks, while schedules differ by country and operator.

Coverage should therefore be verified by country, technology, operator and actual connectivity profile before deployment and major expansion.

Belief Three: The SLA in the Contract Covers the Device

Most connectivity SLAs cover platform availability, meaning the management portal and API. They do not necessarily guarantee that a particular device remains connected to a specific radio network.

For enterprise and industrial deployments, the more important documents are the security and architecture specifications. Key requirements include:

• Private APN: Keeps device traffic away from the public internet and routes it to a defined endpoint.
• Addressing: Static or dynamic private addressing affects firewall rules, device management and incident response.
• Tunneling: IPsec, GRE or SSL tunnels should have defined failover and ownership.
• Data sovereignty: The location where traffic exits the operator network affects jurisdiction and regulatory requirements.
• Scale behavior: Bulk provisioning, pooled data, API limits and billing processes need to work at thousands of devices, not just dozens.
• Support escalation: The contract should explain how faults are escalated when the problem sits with a host operator’s radio network.

Belief Four: Satellite Connectivity Is a Separate Project

Until recently, satellite IoT generally required separate hardware, suppliers and management systems. Standardization is changing that.

3GPP introduced NB-IoT and eMTC support for non-terrestrial networks in Release 17 and completed core and performance requirements for NB-IoT and eMTC over NTN in Release 18. 3GPP NTN overview

Satellite still has important limitations, including higher latency, smaller message sizes, higher cost per byte, developing commercial coverage and antenna constraints. It is not a universal replacement for terrestrial cellular.

The important change is architectural: satellite can increasingly be planned as a fallback layer within the same device strategy rather than as an entirely separate program.

Where Requirements Actually Diverge: Deployment Pattern

Technology comparison tables can hide major differences between deployment types.

Asset tracking and logistics: Devices cross borders, so roaming and mobility matter more than throughput. LTE-M is often appropriate, although the power budget may be dominated by location services rather than the radio.

Smart metering and utilities: Devices are stationary, often buried or enclosed, and may operate for 10–15 years. NB-IoT can fit well, but network longevity and provisioning become major risks.

Industrial monitoring and SCADA: Traffic typically needs predictable access to a plant network without public internet exposure. Private APN, addressing and tunneling can therefore matter more than radio selection.

Fleet and video telematics: High throughput and vehicle power make 4G or 5G appropriate. Here, pooled data allowances and commercial terms can have a major effect on total cost.

Belief Five: The SIM Is a Commodity You Can Decide Last

The SIM is often one of the least reversible parts of an enterprise IoT deployment. It affects the device’s physical design, operator portability, roaming behavior and long-term reliability.

Three decisions belong together:

• Form factor: Removable 2FF, 3FF, 4FF or Tri-cut SIMs allow physical replacement but introduce a socket that can be vulnerable to vibration, corrosion and environmental conditions. Embedded MFF2 removes that mechanical failure point but also removes field replacement.
• Profile management: Multi-IMSI allows one physical SIM to use multiple operator identities. eUICC allows profiles to be downloaded and switched remotely. The GSMA’s SGP.32 specification, published in version 1.2 in June 2024, was designed specifically for constrained IoT deployments. Trusted Connectivity Alliance SGP.32 specification
• Network security and management: The APN, addressing, VPN or tunnel, and connectivity management platform should be considered together.

Specialist providers can package these requirements into one service. Trafalgar Wireless M2M SIM cards: Trafalgar Wireless, for example, supplies removable and embedded MFF2 eUICC M2M SIM options, multi-IMSI carrier selection across 2G, 3G, 4G/LTE, 5G, LTE-M and NB-IoT, private APN and IPsec VPN options, pooled data plans and centralized IoT management.

This is not the only workable procurement model, but combining SIM, profile management and network security decisions can prevent the three areas from being designed independently.

Where This Is Heading

Three developments point toward greater control at the connectivity-management layer.

First, IoT eSIM and eUICC are becoming increasingly important because they reduce the need to lock devices to one operator profile at manufacture.

Second, 5G RedCap is filling the space between low-power cellular technologies and full 5G, targeting devices such as cameras, wearables and industrial gateways.

Third, enterprises with thousands of devices increasingly want one connectivity management platform and API rather than multiple operator portals.

These developments have something in common: they concern who controls operator profiles, addressing and device data rather than which radio technology is selected.

Key Takeaways

1. Radio technology is relatively easy to revise during hardware refreshes. SIM design, provisioning and IP architecture are harder to change.
2. Technical coverage does not guarantee commercial availability. Check country, operator, technology and roaming restrictions.
3. Platform availability in an SLA does not necessarily mean device connectivity.
4. Satellite is increasingly becoming a potential fallback layer within standard IoT architectures.
5. Deployment patterns create different priorities: tracking emphasizes roaming, metering emphasizes longevity, industrial monitoring emphasizes network architecture, and telematics emphasizes data usage.
6. SIM, profile management and network security should be treated as one architectural decision.

Closing Thought

IoT connectivity is often presented as a radio technology comparison because those comparisons are easy to publish. In real deployments, the difficult problems usually involve roaming restrictions, contracts, SIM hardware, network architecture and management systems.

Those are control-layer decisions, and most are made before the first device ships. The decisions that cannot easily be undone once hardware reaches the field should therefore be addressed earlier than the radio comparison itself.