2G/3G Sunset Survival Guide: Your Migration Options Ranked

Who This Is For
You have IoT devices deployed on 2G or 3G networks. You’ve watched the North American and Australian shutdowns from a distance, but now your region is next. You need to understand your options—fast.
This guide ranks migration paths from lowest to highest disruption. By the end, you’ll know which approach fits your fleet, timeline, and budget.
We’ll use a running example: A fleet of 50,000 GPS asset trackers deployed across the UK on O2’s 3G network, currently using a Bluetooth Low Energy (BLE) chip for local configuration. Shutdown deadline: end of 2025.
The Cautionary Tales
If you think carrier shutdown timelines are flexible, look at what already happened.
North America moved fast and didn’t look back. AT&T shut down 3G in February 2022. Verizon followed in December 2022. T-Mobile completed its 2G shutdown in April 2024. Companies that waited until the last minute faced emergency hardware swaps, expedited shipping costs, and angry customers with bricked devices.
Australia completed its 3G shutdown in 2024. Telstra, Optus, and Vodafone all followed through on schedule. Australian 2G has been gone since 2018.
The lesson: carriers don’t extend deadlines. When they announce a date, plan for it.
Where We Are Now
United Kingdom — Shutdown In Progress
The UK 3G shutdown is nearly complete:
- Vodafone: Completed 3G switch-off February 2024
- EE: Completed 3G switch-off January-February 2024
- Three: Completed 3G switch-off late 2024
- Virgin Media O2: Completing 3G shutdown by end of 2025 (started April 2025 in Durham, rolling out regionally)
For our asset tracker fleet: Running on O2 means we have until end of 2025. That’s months, not years.
2G remains available on EE and Vodafone until later in the decade (EE targeting 2029-2030, Vodafone during 2030), but the government mandate requires all 2G to be retired by 2033.
Europe — 2025-2030 Wave
The European shutdown is staggered by country and operator:
- Sweden, Denmark, Austria: Completing 3G shutdowns in 2025
- Netherlands: 2026
- France: 2G until end of 2026, 3G until end of 2029
- Germany: 3G already gone (2021), 2G shutdown by 2028
- Orange Group: Targeting full 2G/3G shutdown across all European markets by 2030
New Zealand — End of 2025
All three New Zealand operators are shutting down both 2G and 3G:
- 2degrees, Spark, One NZ: 3G shutdown end of 2025 into early 2026
- 2G: Also shutting down end of 2025
New Zealand is moving faster than most—both legacy networks disappearing simultaneously. If you have devices deployed there, the clock is ticking.
South Africa — Timeline Uncertain
South Africa’s situation is more fluid:
- MTN: Targeting 3G shutdown December 2025, with a pilot underway in Cape Town
- Vodacom: No formal announcement yet
- Government: Conducting economic impact assessment; full shutdown possibly delayed to 2027
If you’re operating in South Africa, you have more time—but not certainty. Plan for 2025-2027 and monitor carrier announcements.
Migration Options Ranked
From lowest disruption to highest:
1. Bluetooth-Based Connectivity — Lowest Disruption
What it is: Use the existing Bluetooth chip in your device to connect via BLE networks—either crowdsourced smartphone networks or satellite-based infrastructure.
For our asset tracker fleet: These devices already have a BLE chip for local configuration. That chip can become the new WAN link without touching hardware.
| Factor | Assessment |
|---|---|
| Hardware change | None (if Bluetooth chip exists) |
| Firmware change | Yes—new connectivity stack |
| Timeline | Weeks to months |
| Certification | None required |
| Per-device cost | Lowest |
What the migration actually involves:
Firmware integration (2-4 weeks): Integrate a BLE connectivity SDK. Your device already knows how to advertise via Bluetooth for local config—now it advertises location and sensor data for network pickup.
Testing (2-4 weeks): Validate that the new firmware works reliably, doesn’t interfere with existing BLE functionality, and meets your update frequency requirements.
Staged rollout (2-8 weeks): Push firmware OTA to a pilot group, validate in production, then roll out fleet-wide.
No hardware swap. No carrier certification. No manufacturing changes. For our 50,000 trackers, this is a firmware update—not a hardware program.
Tradeoffs: Higher latency (minutes to hours versus seconds), primarily uplink communication. Not suitable for real-time bidirectional control.
Best for: Asset tracking, environmental monitoring, supply chain visibility—any application where you need to know where something is or what condition it’s in, but don’t need instant updates.
Key insight: Many 2G/3G-era IoT devices already have Bluetooth chips that are barely used. Check your hardware specs. If you have Bluetooth, you may have a migration path that requires zero hardware changes and zero certification.
2. LTE-M Module Swap — Moderate Disruption
What it is: Replace the 2G/3G cellular module with an LTE-M module.
For our asset tracker fleet: This means a new module, likely new antenna design, possibly power management changes. Full hardware revision required.
| Factor | Assessment |
|---|---|
| Hardware change | Module + antenna + possibly power system |
| Firmware change | Yes—different AT commands, power management |
| Timeline | 6-12 months |
| Certification | PTCRB (North America), carrier-specific, regional |
| Per-device cost | $15-25 module + certification amortized |
What the timeline actually looks like:
Hardware design (6-10 weeks): Select module, redesign PCB for new footprint, update antenna matching, possibly resize battery.
Prototyping and validation (4-8 weeks): Build prototypes, test RF performance, validate power consumption.
Certification (6-12 weeks): PTCRB, carrier-specific certification, regional certifications as needed. Lab fees run $20K-100K.
Manufacturing retooling (4-8 weeks): Update production line, new test fixtures, updated QA procedures.
Production and deployment (4-12 weeks): Manufacturing run, logistics, field deployment or swap.
That’s 6-12 months best case, assuming no certification failures or design iterations.
For our 50,000 trackers: Module costs alone are $750K-$1.25M. Add certification ($50K+), engineering ($100K-200K in loaded labor), and manufacturing retooling. Total program cost: $1-2M minimum. And we’d miss the O2 deadline starting from today.
Tradeoffs: Still cellular—you’re signing up for ongoing carrier costs and potential future sunset risk (though LTE-M should be safe for 10+ years).
Best for: Applications requiring real-time bidirectional communication, guaranteed uplink, or low-latency alerts.
See The True Cost of Cellular IoT for the full breakdown.
3. NB-IoT Module Swap — Moderate Disruption, Regional Risk
What it is: Replace with an NB-IoT module instead of LTE-M.
For our asset tracker fleet: Similar hardware effort to LTE-M, but NB-IoT doesn’t support handoff between cells—problematic for moving assets.
| Factor | Assessment |
|---|---|
| Hardware change | Module + antenna |
| Firmware change | Yes |
| Timeline | 6-12 months |
| Certification | Similar to LTE-M |
| Per-device cost | Similar to LTE-M |
Tradeoffs: NB-IoT doesn’t support mobility—devices can’t hand off between cell towers while moving. AT&T has deprioritized NB-IoT in North America. Carrier commitment varies significantly by region.
Best for: Stationary devices with very low data needs—smart meters, fixed environmental sensors.
For our asset tracker fleet: Not a good fit. These trackers move. NB-IoT’s lack of mobility support would degrade tracking quality significantly.
Caution: Before choosing NB-IoT, verify your target carriers are committed to long-term support. You don’t want to migrate to NB-IoT only to face another migration in 5-7 years.
4. LTE Cat-1/Cat-1bis Module Swap — Higher Disruption
What it is: Replace with a higher-bandwidth LTE module.
For our asset tracker fleet: Overkill. These devices send 50-byte location payloads. Cat-1 is designed for applications needing 10 Mbps.
| Factor | Assessment |
|---|---|
| Hardware change | Larger module, more power hungry |
| Firmware change | Yes |
| Timeline | 6-12 months |
| Certification | Yes |
| Per-device cost | Higher ($20-40 modules) |
Tradeoffs: Overkill for most IoT applications. Higher power consumption means larger batteries or shorter battery life.
Best for: Applications requiring higher throughput—video transmission, large file transfers, or devices with external power.
For our asset tracker fleet: Would require larger enclosures for bigger batteries. Doesn’t make sense for the use case.
5. Full Platform Redesign — Highest Disruption
What it is: Design a new hardware platform from scratch with modern connectivity built in.
For our asset tracker fleet: Start over. New PCB, new enclosure, new firmware, full certification suite.
| Factor | Assessment |
|---|---|
| Hardware change | Complete |
| Firmware change | Complete |
| Timeline | 12-24 months |
| Certification | Full suite |
| Per-device cost | Highest NRE, but opportunity to optimize |
Tradeoffs: Maximum disruption, longest timeline, highest upfront cost.
Best for: Products that were due for a refresh anyway, or where migration cost exceeds the cost of redesign.
For our asset tracker fleet: If these trackers are 5+ years old and we were planning a v2 anyway, the sunset could be the forcing function. Otherwise, this is the most expensive path.
Silver lining: If you’re going to redesign, you can build in multiple connectivity options (LTE-M + Bluetooth) for future flexibility.
6. Sunset the Product — Exit Option
What it is: End-of-life the product line rather than migrate.
For our asset tracker fleet: Tell 50,000 customers their devices are end-of-life. Refund or replace under warranty if applicable. Exit the market.
| Factor | Assessment |
|---|---|
| Hardware change | N/A |
| Timeline | Immediate |
| Cost | Lost revenue, customer churn, potential contractual issues |
Best for: Low-margin products where migration cost exceeds remaining lifetime value.
This isn’t failure—it’s rational decision-making. Some products aren’t worth migrating.
For our asset tracker fleet: Only makes sense if the product line is already unprofitable or contracts are ending anyway.
Comparison Summary
| Option | Hardware Change | Timeline | Relative Cost | Future-Proof |
|---|---|---|---|---|
| Bluetooth-based | None (if chip exists) | Weeks | Lowest | Yes (unlicensed spectrum) |
| LTE-M swap | Module + antenna | 6-12 mo | Moderate | 10+ years |
| NB-IoT swap | Module + antenna | 6-12 mo | Moderate | Varies by region |
| Cat-1 swap | Module + power | 6-12 mo | Higher | 10+ years |
| Full redesign | Complete | 12-24 mo | Highest | Depends on choices |
| Sunset | N/A | Immediate | Revenue loss | N/A |
Decision Framework
Does your device already have a Bluetooth chip? → Evaluate Bluetooth-based migration first. It’s your fastest, cheapest path if your application tolerates latency.
Do you need real-time bidirectional communication? → LTE-M is your path. Budget accordingly.
Are your devices stationary with minimal data needs? → NB-IoT may work, but verify regional carrier commitment before committing.
Is the product near end-of-life anyway? → Consider sunset, or use this as the forcing function for a full redesign.
How many devices do you have in the field? → Higher counts favor lower-disruption options. At 50,000 devices, the difference between Bluetooth migration and LTE-M swap is over $1 million.
What To Do Next
Audit your fleet. How many devices? What cellular connectivity today? Do they have Bluetooth chips? Pull your hardware specs and BOMs.
Confirm your shutdown timeline. Contact your carriers directly. Published dates are generally reliable, but get confirmation for your specific network and region.
Evaluate latency requirements honestly. Many applications assume they need real-time data but actually tolerate minutes or hours of delay. Challenge your assumptions.
Run the cost model. Compare Bluetooth migration against cellular swap for your specific fleet size. The math may surprise you.
Start now. If you’re in the UK on O2, you have months. If you’re in New Zealand, the same. A 6-12 month LTE-M migration timeline means you’re already too late if you haven’t started.
For our 50,000 UK trackers: The answer is clear. These devices have Bluetooth hardware. The application tolerates minutes of latency. Bluetooth-based migration can be completed in weeks—well before O2’s deadline—at a fraction of the cost of an LTE-M swap. The $1M+ we’d spend on cellular migration stays in the bank.
The companies that navigated North American and Australian shutdowns successfully started early and evaluated all options—not just the obvious cellular-to-cellular swap. Many discovered they’d been over-specifying connectivity for years, paying for real-time capability they never actually used.
Hubble enables global connectivity using Bluetooth LE hardware. No module swap, no certification. Start for free and pay-for-usage pricing. See how it works →