Every cellular IoT device carries a small piece of hardware that quietly decides who it does business with. The SIM holds the credentials that let a meter, a truck tracker or a payment terminal join a mobile network. For most of the history of machine connectivity, that SIM also tied the device to one provider for as long as it stayed in the field.
For a phone, that is a minor inconvenience. People change phones every few years and can swap a SIM card in seconds. For an industrial device bolted inside a cabinet, buried in a meter pit or mounted on a vehicle in another country, changing the SIM means sending a technician. Multiply that by thousands of devices and the cost of switching providers can outweigh any saving from a better contract.
A newer standard from the GSMA, the mobile industry’s trade body, is meant to change that. It is called SGP.32, and in September 2026 it was at the centre of several product launches and a lively argument about who should control the connection inside a connected device.
Why The SIM Became A Lock
To see why SGP.32 matters, it helps to look at how IoT connectivity has traditionally been sold.
One Device, One Operator
A company building a connected product usually signs a connectivity agreement early in development. The SIMs that go into the first production run belong to that provider. The devices are tested, certified and shipped with that provider’s credentials inside.
That arrangement can hold for a decade or more. Industrial devices, smart meters and vehicles are often designed to stay in service for ten years or longer. Over that time, prices change, coverage changes, networks are retired and business needs shift. The connectivity contract signed at launch may no longer be the right one years later, but the SIM inside every device still points to the original provider. We compared the main connection options for industrial sites in our industrial connectivity explainer.
The Cost Of Switching In The Field
In theory, a company can move to a new provider by replacing the SIMs. In practice, that means reaching every device physically. For assets spread across a country or several continents, the labour and logistics can run far beyond the cost of the connectivity itself.
The problem sharpens when networks are switched off. As operators retire older 2G and 3G networks, devices tied to one provider can lose service entirely. A recent piece from connectivity provider Eseye on why UK networks are being switched off recommended SGP.32-compliant eSIM platforms as a way to move devices to other networks over the air, without physical SIM swaps.
How eSIM Tried To Fix It
eSIM, or embedded SIM, replaces the fixed credentials of a traditional SIM with a chip, known as an eUICC, that can hold one or more downloadable operator profiles. In principle, that allows a device to change provider by downloading a new profile rather than swapping hardware. In practice, the rules for how that download happens have shaped who benefits.
SGP.02: Built For Operators
The GSMA’s first remote SIM provisioning specification for machines, known as SGP.02, was designed around mobile operators. Profile changes were pushed from network-side systems, and switching providers generally required cooperation between the old operator and the new one. For many enterprises, that meant the flexibility existed on paper but was difficult to use when commercial interests pointed the other way.
SGP.22: Built For Phones
A later specification, SGP.22, was designed for consumer devices such as smartphones and smartwatches. It lets the user choose and download a profile, typically by scanning a code or using an app on the device. That works when a person is holding the phone. It does not suit a sensor with no screen, no keyboard and no user.
SGP.32: Built For IoT
SGP.32 was written specifically for IoT devices. It keeps the idea that the customer, rather than the operator, chooses the profile, but adapts it for devices that are unattended and often low-powered.
Two components sit at the heart of it. The first is the eSIM IoT remote Manager, or eIM, a system that lets the device owner manage profiles across a fleet remotely: downloading, enabling, disabling and deleting them. The second is the IoT Profile Assistant, or IPA, a small function on the device or the eUICC that carries out those instructions. Together, they are meant to give the company that owns the devices, rather than any single network provider, control over which network each device uses.
What Launched This Month
Two announcements in September show how providers are turning the standard into products, and how differently they are approaching it.
Onomondo’s Remote Manager
On 16 September, Danish connectivity provider Onomondo launched its own eIM. According to IoT Business News, the service lets companies download, enable, disable and delete eSIM profiles on devices already deployed, without physical access.
The notable detail is that Onomondo says the remote manager works independently of its own connectivity. Customers can use it to manage profiles from other providers, or as part of Onomondo’s wider SGP.32 offering, which includes eUICC SIMs, profiles and global connectivity. The company is targeting fleets expected to stay in service for ten years or more, including vehicles, industrial machinery, meters and payment terminals.
Henrik Aagaard, Onomondo’s chief technology officer, framed the launch as a shift in control: “SGP.32 is the industry’s opportunity to finally put IoT companies in control of their cellular connectivity.”
Granite’s Three-Carrier SIM
A day earlier, US communications provider Granite launched SIM360, a platform that places connectivity from all three major US wireless carriers on a single SIM, managed through one interface. IoT Business News reported that the platform uses the SGP.32 architecture and was developed with Thales.
Granite is aiming SIM360 at higher-bandwidth IoT and fixed wireless access deployments, connected fleets, logistics, remote offices, kiosks and field equipment. It is available now for domestic US deployments. Granite’s president and chief executive, Rob Hale, said the product brings “all three major U.S. wireless networks together on a single SIM.” Guillaume Lafaix of Thales described the aim as “greater resilience and flexibility, without adding complexity for businesses.”
The two launches illustrate different uses of the same standard. Onomondo’s emphasis is on letting customers move between providers over a device’s life. Granite’s emphasis is on resilience across several networks within one market.
Who Benefits Most
SGP.32 is not equally valuable for every deployment. Its benefits concentrate in a few situations.
Devices That Outlive Their Contracts
The clearest case is long-lived equipment. A device expected to run for ten or fifteen years will almost certainly outlast its first connectivity contract, and possibly the network technology it was launched on. Being able to change profiles remotely turns connectivity from a decision made once at launch into one that can be revisited.
This is also where the economics of lock-in bite hardest. When switching requires a truck roll to every site, the incumbent provider has considerable bargaining power at renewal time. Remote profile management weakens that position.
Markets That Restrict Roaming
The second case is global deployments. Many IoT devices rely on roaming, connecting through a home provider’s agreements with local networks. Some countries restrict permanent roaming, requiring devices that stay in the country to use a local network.
In an interview with IoT Business News, Onomondo co-founder Michael Karlsen pointed to Brazil, Turkey, China and Saudi Arabia as examples of markets where provider flexibility matters most. Being able to download a local profile after a device arrives in the country can make the difference between a compliant deployment and one that has to be redesigned.
What SGP.32 Doesn’t Solve
The standard has attracted enthusiasm, but it is not a simple fix, and some of its supporters say so openly.
Added Complexity
In the same interview, Karlsen described SGP.32 as “not a panacea.” Managing multiple profiles, providers and contracts adds operational work. Devices need firmware that supports the standard, and profiles from different operators must be tested and approved for the device. Karlsen recommended that companies weigh the risks and benefits before adopting it, noting that short-lived or purely local deployments may not need the flexibility.
There is also the question of existing fleets. Devices already in the field with older SIMs or earlier eSIM standards will not gain SGP.32 capabilities without hardware or firmware changes. For many companies, the benefits will arrive gradually as new device generations replace old ones.
Control Shifts, It Doesn’t Disappear
SGP.32 moves control over profiles towards the device owner, but someone still has to run the eIM. If that remote manager is operated by a single connectivity provider and cannot easily be replaced, a new form of lock-in can appear one layer up.
That is why Onomondo’s decision to offer its eIM independently of its connectivity is notable, and why buyers are likely to pay close attention to how portable each provider’s remote management really is. The standard creates the possibility of choice. Whether customers actually get it depends on how products are built and contracts are written.
What It Means For Mobile Operators
For mobile operators, SGP.32 cuts both ways. On one hand, it weakens a source of stable revenue. When switching required a site visit, customers tended to stay put, and IoT contracts could run for many years with little price pressure. If devices can move with a remote instruction, operators will have to compete harder to keep them.
On the other hand, the standard can open doors. An operator with strong coverage in a particular country can win devices that were originally connected elsewhere, simply by providing a profile that the customer downloads. In markets that restrict permanent roaming, a local operator may gain business from global device makers that previously relied on roaming agreements. Operators that make their profiles easy to test, approve and download could find that flexibility works in their favour.
The balance will depend on how many devices adopt the standard and how actively customers use the ability to switch. Many companies may keep their original provider for years and value SGP.32 mainly as insurance: a guarantee that they could move if prices rose, service declined or a network closed. Even unused, that option changes the negotiation.
Questions Buyers Should Ask
For companies planning new cellular IoT deployments, SGP.32 is worth raising early in design and procurement. A few questions help separate real flexibility from marketing.
Does the device hardware and firmware support SGP.32, including the IoT Profile Assistant? Retrofitting later is far harder than designing it in.
Who operates the eIM, and can it be replaced without replacing the devices? A remote manager tied to one provider may recreate the lock-in the standard is meant to remove.
Which operator profiles have been tested on the device, in which countries? Theoretical access to many networks is less useful than proven access to the ones you need.
What does a switch actually cost? Moving profiles may be technically simple but still carry contract, testing and support costs.
How long is the device expected to stay in service? For a two-year deployment in one country, a traditional SIM may be perfectly adequate.
Connectivity As A Choice
For most of its history, cellular IoT has treated connectivity as something decided once, at the start, and then lived with. The SIM inside the device made that decision hard to undo.
SGP.32 does not remove every obstacle, and its adoption will take years as new devices replace old ones. But it changes the default. If devices can move between networks remotely, connectivity becomes something companies can revisit as prices, coverage and regulations change, rather than a commitment sealed into the hardware.
Karlsen’s view is that within three to five years, connectivity could become “a point of zero friction in building IoT.” That is a forecast, not a certainty. The September launches suggest the industry is at least starting to build the tools that would make it possible.
