SGP.32 Support

Built for the SGP.32 era

SGP.32 is the GSMA eSIM standard built specifically for IoT. Published in May 2023, it introduced the eIM (eSIM IoT Manager) role, the IPA (IoT Profile Assistant), and CoAP-based provisioning for constrained devices on NB-IoT and LTE-M. It is the first RSP standard genuinely designed for headless IoT at scale.

eSIM IoT Manager provides the management layer above the SGP.32 standard – the operational platform that enterprise teams need to actually run SGP.32 deployments without taking on unsustainable in-house complexity.

New to SGP.32? The most complete reference for the standard is sgp32.co.uk, an independent UK reference site covering the architecture, hardware landscape, and implementation considerations in depth.

What SGP.32 changed

Before SGP.32, enterprises managing IoT eSIM had two options: SGP.02, which required SM-SR infrastructure and created operator lock-in, or SGP.22, which required a user interface present to initiate profile changes. Neither was built for the reality of mass IoT deployment.

SGP.32 addressed both problems by introducing server-initiated provisioning (the device does not need to pull), an eIM architecture that removes SM-SR lock-in, and CoAP transport that works within the data budget constraints of NB-IoT sensors.

What SGP.32 does not solve on its own

SGP.32 is a technical specification for profile delivery. It does not define:

These are not gaps in the standard – they are enterprise operational requirements that sit above it. eSIM IoT Manager addresses them.

SGP.32 and the DIY risk

Enterprises managing SGP.32 without a managed platform take on significant operational complexity. Profile management across multiple operators, eIM configurations, bootstrap provisioning, and network policy management each require sustained technical resource. The more operators and device types involved, the more this multiplies.

The risk is not that SGP.32 is difficult – the standard is well-designed. The risk is that operational complexity is easy to underestimate at the proof-of-concept stage and expensive to discover in production with 50,000 devices in the field.