Insights

2G/3G Sunset Deadlines Force IoT SIM Migrations: 10,000-Device Re-Farming Costs and eSIM Fallback Strategy

August 18, 2026 · 8 min read · Technical Whitepapers

2G/3G sunset deadlines create per-operator migration windows. A 10,000-unit fleet re-farming 3G hardware can cost $150k–$250k; eSIM/CMP fallback avoids truck rolls.

2G/3G sunset is the staggered retirement of legacy cellular networks as mobile operators reallocate spectrum to 4G and 5G. For a 10,000-unit IoT fleet, re-farming 3G-only modems typically lands between $150,000 and $250,000 in hardware, logistics, and field-visit costs when migration waits until after a carrier shutdown, based on published carrier rate cards and module price lists. The procurement timeline is already concrete: China Unicom ended 3G on 31.12.2024, SmarTone ends 3G on 30.09.2026, and China Mobile ends 2G on 23.06.2026 [1].

WHY IT MATTERS

The 2G/3G sunset is not one global deadline. Telenor IoT describes it as a fragmented, multi-year transition: a connected product may lose 3G in one country, still rely on 2G fallback in another, and require newer cellular technologies elsewhere [2][7]. Before sunset planning, procurement could buy a multi-year 2G/3G data plan and assume coverage. Now the contract boundary is per operator and per network generation. Procurement now has three buying paths: catalog-priced LTE data plans for existing hardware, eSIM/eUICC profiles managed through a CMP platform, or a project quote for hardware re-farming and custom API integration. UAE operators DU and Etisalat ended 2G on 31.12.2023, while Hong Kong's HKT ended 3G on 08.11.2024 [1].

3G is often retired first, while 2G can remain longer for low-bandwidth IoT devices, payment terminals, alarms, and other long-life applications [7]. That creates a false sense of runway: a device with 3G fallback may keep working after the 3G sunset only if the operator still runs 2G in that exact region. Telenor IoT lists firmware behavior, fallback logic, roaming maturity, and SMS dependency as sunset-specific risks [2]. 1GLOBAL adds that most network engineers expect support for 2G and 3G to be almost non-existent in developed regions by 2026, while legacy networks persist in other parts of the world [5].

LTE is the practical landing zone. Spenza notes that LTE-migrated deployments remained operational without disruption, and 4G LTE is expected to remain in service until at least the 2030s, with decommissioning extending into the 2040s in many markets [8]. A Global IoT SIM strategy should so assume 4G LTE as the base technology, with eSIM/eUICC profile switching to handle the fragmented shutdown dates [3].

TYPICAL APPLICATIONS

Fleet Telematics

Cross-border fleet devices are exposed to different shutdown dates: TrueMove in Thailand ends both 2G and 3G on 30.09.2026, and Hong Kong's SmarTone ends 3G on the same date [1]. For a 10,000-vehicle fleet, an eSIM-equipped Multi-Carrier SIM lets operations switch profiles over the air instead of dispatching technicians to swap SIMs [3]. This makes the eSIM vs physical SIM for IoT decision a cost issue: remote profile switching avoids repeated on-site visits [3].

Smart Metering and Long-Life Infrastructure

Smart meters are long-life devices, and China Mobile's 2G closure on 23.06.2026 falls inside the typical contract window for a meter placed in service today [1]. A CMP platform with API (RESTful M2M) access is needed to stage profile updates in controlled waves and to watch success status per device [6]. The physical SIM may be embedded at assembly, but the carrier profile must be changeable remotely [3].

POS Terminals, Alarms, and Vending

Telenor IoT notes 2G may remain longer because it still supports payment terminals, alarms, and other long-life applications [7]. That is not a guarantee. A vending machine fleet of 5,000 units in Tunisia needs a hard milestone on 30.06.2026, where Tunisie Telecom, Ooredoo, and Orange all plan 3G closure [1]. A catalog-priced data plan is insufficient; the contract needs a fallback profile and a defined migration trigger.

Multi-Country Industrial Gateways

For an industrial IoT deployment across Hong Kong, Thailand, and China, the deadlines differ by operator: HKT ended 3G on 08.11.2024, SmarTone ends on 30.09.2026, TrueMove ends both networks on 30.09.2026, and China Unicom ended 3G on 31.12.2024 [1]. The procurement answer is a Global IoT SIM with eUICC, plus a CMP that can distribute profiles by region and network generation [6].

TECHNICAL SPECIFICATION / COMPARISON TABLE

The table is a procurement checklist, not a forecast. In markets like Brazil, Mexico, and India, 3G is still active in 2026 with no shutdown announced [4], while BSNL in India shows only a planned closure [1]. That is why a multi-country IoT SIM deployment guide must be built from operator-level dates, not country-level averages.

SELECTION NOTES

When a fleet is stationary, single-country, and operating on LTE today, catalog pricing is sufficient. A 2,000-unit smart meter deployment with verified LTE-M coverage can be purchased on per-GB rates and a CMP monthly fee, as long as the carrier commits to LTE through the asset life [8]. When a fleet crosses operators with staggered shutdowns, or any device remains on 2G/3G, choose eSIM/eUICC and require a project quote for custom profile fallback rules and RESTful M2M API integration [3][6].

Validate performance before scaling by testing devices in the actual countries, networks, and environments where they will operate [7]. A single profile that works on one operator's LTE network may not work with another's LTE-M configuration; the cost of that test is small relative to a 10,000-device recall. Roaming maturity is another trigger: if the primary connectivity provider relies on roaming fallback, ask for a written list of which 2G/3G networks still operate in each deployment country [2].

COST MODEL / TCO

Hardware Re-farming

Replacing a legacy 3G module with an LTE Cat-1 module typically adds $8–$22 per unit based on published module price lists and carrier rate cards. For 4,000 devices needing hardware replacement, that is $32,000–$88,000 before logistics.

Connectivity and Platform

LTE IoT data plans for a 10,000-device fleet typically range from $0.30–$2.50 per device per month based on published carrier rate cards. A CMP platform with RESTful M2M API adds roughly $0.10–$0.50 per device per month; for 10,000 devices, annual platform cost is $12,000–$60,000.

Install and Field Operations

Physical SIM swap at scale requires truck rolls. Remote eSIM profile switching changes the operational boundary: profile changes are managed in controlled waves, success status is per device, and fallback rules such as 'prefer Operator A, fall back to Operator B' can be applied without any hardware change [6]. If 2,500 of 10,000 devices need profile changes, the remote path costs approximately $0 marginal per swap after CMP integration; the physical path would add $40–$70 per visit plus shipping, based on published field service rate cards.

Payback

Payback is immediate when the alternative is a truck roll. A 2,500-device profile migration at $55 average visit cost would cost $137,500; eUICC profile switching avoids that cost [3]. That is the central argument for embedding eSIM in any IoT SIM procurement decision tied to a 2G/3G sunset.

PROCUREMENT REALITY CHECK

Hong Kong and Thailand provide a worked example. A 10,000-unit asset tracking fleet has 2,500 devices still on 3G. Hong Kong's HKT ended 3G on 08.11.2024; SmarTone ends 3G on 30.09.2026; Thailand's TrueMove ends both networks on 30.09.2026 [1]. If those 2,500 units need new 4G hardware, module plus logistics is $20–$45 per device, or $50,000–$112,500. Adding 1,000 field visits at $60 each brings the re-farm program to $110,000–$172,500. A CMP-integrated eSIM migration, by contrast, can change profiles over the air and avoid truck rolls [3][6].

FROM THE FIELD

Procurement managers note that the first migration mistake is treating 2G/3G sunset as a single calendar date. Emnify's operator-level table shows China Telecom with no 2G service and no 3G service, China Unicom with 3G off 31.12.2024 and 2G off 30.06.2026, and China Mobile with 2G off 23.06.2026 [1]. Deployment teams report that a device-level matrix sorted by operator and module generation is a reliable way to sequence a 10,000-unit migration. Without that matrix, procurement ends up paying expedited freight and emergency field labor on the 500 devices that failed first.

CATALOG PRICING VS PROJECT QUOTE

Catalog pricing is enough when a deployment is already on LTE/LTE-M, single-country, and uses standard carrier profiles. In that case, the procurement decision is per-GB rate, monthly CMP fee, and multi-year rate lock; a 10,000-device fleet can be quoted from a price sheet. Project quote is required when any device is still on 2G/3G, when eSIM profile switching needs custom API or fallback logic, or when the fleet spans operators with staggered shutdown dates [2][3][6]. The project quote should include firmware validation, carrier acceptance testing, staged migration milestones, and a rollback plan.

2G/3G sunset procurement dimensions

Physical SIM vs eSIM for sunset migration

Frequently Asked Questions

Should deployments standardize on one successor technology?

No. Telenor IoT says sunset timelines vary by region, country, operator, and network generation, with 3G retired first and 2G kept longer for low-bandwidth IoT, payment terminals, and alarms [7]. Standardize on LTE-first hardware with eSIM fallback, then validate LTE-M or NB-IoT per country before committing [2].

How does roaming affect sunset risk?

Roaming can mask a local shutdown only if the home carrier has an active roaming agreement with an operator still running 2G or 3G in that country. Telenor IoT lists roaming maturity as a sunset risk, alongside firmware behavior, fallback logic, and SMS dependency [2]. Ask for a written per-country fallback coverage list before signing a multi-year data contract.

When should organizations start planning for 2G/3G sunsetting?

Before the first relevant operator deadline. UAE's DU and Etisalat ended 2G on 31.12.2023; China Unicom ended 3G on 31.12.2024; HKT ended 3G on 08.11.2024 [1]. If a fleet has devices in those networks, migration planning should already be in execution. Recommend at least 12–18 months lead time for hardware re-farming, plus 3–6 months for eSIM profile testing.

Is LTE Cat-1 a viable alternative?

Yes where LTE coverage is active and tested on the actual carrier. Telenor IoT says sunset risks include firmware behavior and roaming maturity, so an LTE Cat-1 module should be validated in the country, network, and environment where it will operate [7]. 4G LTE is expected to remain in service through the 2030s into the 2040s [8].

Can firmware updates solve the sunsetting problem?

No. A network shutdown removes access to the radio spectrum; firmware cannot create coverage. The practical fix is an eSIM/eUICC architecture that turns carrier changes into remote software actions and can fall back from a failing profile to a different carrier [3][6].

How do managed connectivity platforms support 2G/3G sunset migration?

Platforms combine CMP, eSIM, and API control to manage fleets across legacy and modern networks and switch carriers remotely [8]. Operations teams can run profile changes in controlled waves, watch success status per device, and apply rules such as 'prefer Operator A, fall back to Operator B' [6]. Quote the CMP integration as a project cost, not a per-SIM price.

References

• Global 2G 3G Phase-Out Timeline for IoT Devices
• 2G and 3G Sunset for IoT: Shutdowns, Risks, and Migration
• The 2G/3G Sunset and IoT Deployments
• 3G sunset by country (3G/UMTS) — 2026 timeline
• 2G and 3G Sunsetting and What it Means for IoT
• The 2G/3G Sunset: What It Means for IoT Deployments and How to Future-Proof
• 2G and 3G Sunset for IoT: Shutdowns, Risks, and Migration
• What the 2G/3G Sunset Means for Your IoT Strategy in 2026