Within ten years, the number of cars linked to the cloud is expected to more than double. Omdia's latest connected vehicle forecast puts the global installed base at 1.05 billion vehicles by 2035, compared with 473.7 million in 2025. That is growth of roughly 8.3% a year, sustained for a decade.
The size of the fleet is not the part Omdia chose to underline, though. Its sharper point is about timing. Mobile operators are expected to begin retiring LTE networks from 2030, and a large share of the cars sold between now and then will lean on LTE for things drivers now take for granted: software updates delivered over the air, route planning for electric vehicles, data feeds for driver assistance, and emergency calling.
In Omdia's view, the end of LTE will weigh more heavily on the auto industry than the 2G and 3G shutdowns did. The reason is simple. Those older networks mostly carried add-on services. LTE now carries the connected car's core functions.
"As the industry advances toward software defined vehicles—and increasingly explores what AI defined vehicles might look like—it has rightly embraced foundational enablers such as eSIM and hybrid terrestrial/non terrestrial (TN/NTN) connectivity," said John Canali, Principal Analyst, IoT at Omdia. "However, progress on 5G adoption has been slow. With mobile network operators beginning to decommission LTE from 2030, automakers face a worst case scenario: vehicles that depend on connectivity but cannot attach to a viable network."
What The Omdia Forecast Actually Says
Omdia's model tracks the connected vehicle base by modem type, which is what makes the LTE warning possible. It is less about how many cars will be online than about which radio each of those cars carries, and whether a network will still exist to answer it.
The firm expects the next phase of the market to be shaped by three things: stricter service level agreements (SLAs) between automakers and their connectivity partners, wider use of software updates pushed over the air (OTA), and vehicle software designed around services that can earn money after the sale. All three assume a reliable data link that lasts for the whole life of the car.
That assumption is where the risk sits. A subscription feature, a remote diagnostic or a security patch is only worth something if the vehicle can still reach the network in year eight, ten or twelve of its life.
Why The Forecasts Do Not Agree On Size
Analysts count connected cars differently, and the gap between estimates is wide. Transforma Insights, for instance, projects a connected vehicle fleet of 2.1 billion by 2035, up from 876 million in 2025, with IoT revenue of about USD 72 billion a year by the end of the period. The difference with Omdia comes largely from what each firm counts as a connected vehicle and how it defines its base.
What both firms agree on is direction. The fleet roughly doubles or more, and the radio inside it shifts toward 5G. Transforma expects 5G's share of annual connected vehicle shipments to climb from 11% in 2025 to 93% in 2035, with most of those units being multi-mode 4G/5G designs rather than 5G-only. It also notes that automakers have so far chosen 5G mainly for longevity and bandwidth, not for advanced network features.
Put simply, the shift to 5G is coming, but slowly enough that hundreds of millions of LTE-only vehicles will be on the road when the first LTE networks start to go dark.
Why LTE Is A Bigger Problem Than 3G Was
The industry has been through a network retirement before, and it was not smooth.
What The 3G Shutdown Taught Automakers
When US operators shut down 3G in 2022, owners of relatively recent cars discovered that features they had paid for simply stopped working. Volkswagen, for example, told its dealers in January 2022 that Car-Net vehicles from model years 2014 to 2019 would lose connectivity from 1 February 2022, ahead of AT&T's 3G switch-off. Remedies across the industry varied from brand to brand and from model year to model year: some owners got a software update, some needed a dealer to swap hardware, some were offered a paid upgrade or a plug-in adapter, and some got nothing at all.
The cars kept driving. What disappeared were remote lock and unlock, in-car hotspots, app control, automatic crash notification and, in some cases, the ability to receive software updates over the air.
LTE Now Carries Far More Of The Car
That was a painful episode, but it touched mostly peripheral services. The vehicle of 2026 is built differently. Connectivity is no longer a bolt-on telematics box that handles an SOS button and a remote start. It is how the car receives fixes, security patches and new functions after it leaves the factory, how an electric vehicle plans charging stops, and how driver-assistance systems pull map and traffic data. We traced that shift in more detail in our look at how IoT rewired the car.
If the modem can no longer attach to a network, the car does not just lose convenience features. It loses the channel through which its maker keeps it secure and up to date. For a software-defined vehicle, that can be the difference between a car that improves with age and one that is frozen at its last update.
The Lifecycle Mismatch At The Heart Of It
Cars live much longer than phones, and they are getting older. The average US light vehicle reached 12.8 years of age in 2025, according to S&P Global Mobility, with a scrappage rate of about 4.5% a year. A car sold in 2027 with an LTE-only modem could reasonably still be on the road well into the late 2030s.
Now set that against network plans. Industry estimates compiled by ABI Research suggest LTE sunsets in the United States could arrive between 2030 and 2032, while Europe may keep LTE running until 2035 or later. Those dates fall squarely inside the expected life of vehicles being designed and sold today.
The planning problem is made worse by the way vehicles are built. Electronics for a vehicle program are locked years before launch, and a platform can stay in production for most of a decade. That means modem choices made in 2025 and 2026 decide which radios are still on the road in 2040.
There Is No Single Switch-off Date
Network retirements are not coordinated events. Each operator sets its own timetable, usually driven by the value of refarming spectrum for newer technology, and plans can shift. A car sold into dozens of markets faces dozens of separate schedules, and a single vehicle can cross several of them in one road trip. The same patchwork is playing out with 2G and 3G today, as we covered in our piece on the UK switch-off.
For automakers, this makes LTE's end less a cliff than a slow erosion: coverage thins in one country, then another, until a feature that worked everywhere works only in some places.
Europe's ECall Shows How Messy It Gets
The clearest preview of what an LTE sunset could look like is already unfolding in Europe, with 2G and 3G.
Since 2018, new car and van types sold in the EU have had to carry eCall, the system that automatically dials the 112 emergency number after a serious crash and sends the vehicle's location. The original design relied on 2G and 3G voice calls. A study carried out by LS Telcom for the European Commission, published in May 2026, found that an estimated 64 million vehicles could lose eCall because national network shutdowns are not coordinated. It also found that aftermarket fixes are not expected before 2030.
Regulators have moved new vehicles onto a newer standard. Under Delegated Regulation (EU) 2024/1180, Next Generation eCall over 4G and 5G is required in new vehicle types from 1 January 2026 and in all new vehicles from 1 January 2027.
That fix carries its own lesson. The emergency system is now moving onto packet-switched networks, including LTE, the very generation that operators expect to begin retiring from 2030. Vehicles whose modems can fall back on 5G should be better placed. Those that rely only on LTE could face a similar question a decade from now.
What 5G, RedCap And ESIM Change
Three technologies are usually cited as the way out. Each helps, but none solves the problem on its own.
5G Has Been Slow To Arrive In Cars
Canali's point about slow 5G adoption comes down largely to cost and fit. Full 5G modules are expensive and offer far more capability than a typical telematics unit needs. 5G RedCap (reduced capability) was designed to sit in between, but pricing has been the obstacle: ABI Research noted in early 2025 that RedCap modules carried roughly triple the price of an LTE Cat-4 module.
Suppliers are now working to close that gap. Marelli, for example, launched an affordable 5G RedCap telematics solution in December 2025, pitched explicitly as an alternative to full 5G ahead of 4G phase-outs, with the ability to switch between 4G and 5G networks. The broader trade-off between capability and module cost is one we explored in why Cat-1 bis won in cellular IoT, and the modem itself keeps taking on more work, as our piece on the cellular module becoming a computer describes.
ESIM And Satellite Are Insurance, Not Replacements
eSIM lets an automaker change the operator profile on a car over the air, which is useful if one operator retires LTE faster than another in the same market. But eSIM changes the subscription, not the radio. An LTE-only modem with a new profile is still an LTE-only modem. The wider tussle over who controls those profiles is covered in our explainer on the fight over the IoT SIM.
Hybrid terrestrial and non-terrestrial connectivity adds a satellite path for coverage gaps and emergency messaging. That can keep a safety link alive where cellular coverage fades, but satellite links today carry a fraction of the data a car pulls over a cellular network. Projects to beam standard mobile signals from orbit, like the one described in our look at AST SpaceMobile's space-based towers, are still maturing.
What Automakers Are Being Pushed To Do Now
Omdia's advice is for automakers to match their vehicle roadmaps, connectivity contracts and hardware upgrade plans to the dates mobile operators set for their networks. In practice, that points toward a handful of decisions that need to be made while vehicles are still on the drawing board:
· Choosing modems with 5G support, including multi-mode 4G/5G and RedCap parts, for programs that will stay on the road beyond 2030.
· Writing network lifetime into connectivity contracts, including notice periods and commitments on what happens when a network generation is retired.
· Keeping modem firmware and connectivity software updatable over the air, so vehicles can adapt as networks change.
· Planning retrofit or replacement options early, rather than scrambling in the months before a shutdown, as happened with 3G.
· Being clear with buyers about which connected features depend on which network, and for how long those features are expected to last.
None of this is free. Each choice adds cost to a vehicle at a time when automakers are under pressure on price. The alternative, as the 3G episode showed, is a costlier mix of recalls, dealer visits, refunds and damaged customer trust.
Questions Worth Asking Before Buying
For fleet managers and private buyers, the practical takeaway is to ask about the radio, not just the features. Useful questions include which network generation the vehicle's modem supports, whether it can use 5G, how the automaker plans to handle future network retirements, and which services would stop working if the car could no longer connect. For commercial fleets, the same questions apply to aftermarket telematics devices, which often have shorter design cycles but can end up in service for just as long.
The Bottom Line
Omdia's headline number, a billion-plus connected vehicles by 2035, is a sign of how central connectivity has become to the car. The warning attached to it is a reminder that the network underneath those vehicles has a shorter life than the vehicles themselves.
The 2G and 3G shutdowns cost drivers convenience features. With LTE, the stakes include software updates, security patches and safety services. The automakers that treat connectivity as part of the vehicle's lifespan, and not as a component chosen once and forgotten, are likely to have far fewer difficult conversations with customers in the 2030s.
