Most IoT budgets are built around the thing you can hold in your hand. In a regional IoT spending forecast published in June 2025, IDC put hardware at more than 60% of all IoT spending in 2025 across Asia/Pacific, driven mainly by module and sensor purchases for connecting endpoints to the network. Hardware is also the one line that stops. You buy the device once. The SIM underneath it bills every month for as long as that device stays in the field, and an industrial-grade SIM is built to survive out there for well over a decade.
That asymmetry is the whole problem with IoT connectivity budgeting. The largest number in year one is almost never the largest number across the life of the deployment, and the recurring line is the one that gets approved on a single per-megabyte figure pulled off a competitor’s website.
This guide walks through how enterprise IoT connectivity costs are actually structured: the billing models, the line items that sit outside the headline rate, the cross-border traps, and two worked models you can rebuild with your own numbers. It is written for the person who has to defend a figure in a spreadsheet, not for someone browsing rate cards.
Why This Guide Does Not Publish a Rate Card
A quick note on what you will not find below, because it matters to how you use the rest of the page.
We do not publish per-megabyte list prices for ourselves or for anyone else. Connectivity pricing in this market is quote-based for a reason: the same 10 MB costs different amounts depending on the country, the network agreement behind it, the volume you commit to, the term, the SIM form factor, and whether your traffic crosses a private APN on the way to your cloud. A published number stripped of those variables is not a price, it is a marketing figure, and building a budget on it is how teams end up 40% over plan without ever seeing an unexpected invoice line.
What follows instead is the cost structure, the arithmetic, and the exact questions that force a comparable quote out of any provider. That is more durable than a table of numbers that decays within two quarters.
The Five Things Enterprises Underestimate in an IoT Connectivity Budget
These are the five items that most often separate the approved budget from the actual spend. Each one is worth pricing explicitly before you sign.
Device count drift. The number of devices you budget is rarely the number you get billed for. Spares sitting in a warehouse, units returned under RMA, replacements shipped ahead of failures, and SIMs that were deactivated in the field but never cancelled in the platform all carry a monthly SIM fee in most commercial models. A 5,000-unit deployment routinely bills across 5,300 to 5,600 active lines. Ask how suspended and stored SIMs are charged, and whether there is a pre-activation state that carries no fee until the device first connects.
The average is not the number that sizes the plan. Fleet data use is a distribution, not a mean. A tracker that averages 40 MB a month will include units that idle at 5 MB and units that sit in a congested reporting loop at 300 MB. If your plan is sized on the average and priced per device, the tail devices generate overage. If it is pooled, the tail is absorbed. This single decision moves enterprise IoT connectivity costs more than any negotiated rate.
Firmware and over-the-air updates. A 4 MB firmware image pushed to 5,000 devices is 20 GB of billable traffic in one afternoon. Teams that update quarterly are adding 80 GB a year that never appears in the steady-state model. Budget updates as a separate line with a stated frequency and image size.
Security is a recurring line item, not an included feature. A private APN, a site-to-site VPN, static IP addressing, and IMEI locking are commonly billed as recurring charges by the connection or by the APN. They belong in the monthly model from day one.
The exit is priced too. Minimum terms, early termination fees, and whether you can move your SIM estate to another operator profile without replacing hardware all determine what your renewal looks like. A provider you cannot leave is a provider that can reprice you.
The Three Pricing Models, Compared
Nearly every quote you receive will be one of three shapes, or a hybrid. Understanding IoT connectivity pricing starts with knowing which shape you are looking at, because the same deployment can look cheap in one model and expensive in another.
Per-megabyte (pay as you use) pricing charges you for the data you actually consume, usually as a per-MB rate plus a fixed monthly fee per SIM. It suits deployments with low or unpredictable consumption, where paying for headroom you never use is the bigger waste. The risk sits in the tail: a misbehaving device or a chatty firmware release bills at the full rate with nothing to cap it.
Per-device-per-month (bundled allowance) pricing gives each SIM a fixed data allowance for a fixed monthly price, with an overage rate above it. It is the easiest model to forecast and the easiest to over-buy, because you must size every device for its worst month. Overage rates above the bundle are frequently several times the effective in-bundle rate, which is where the surprise invoices come from.
Pooled (dynamic) data pricing allocates a single data volume to the account rather than to individual devices, and lets each device draw what it needs from the shared pool. Devices that under-use subsidize devices that over-use, and the aggregate is what you buy. Pooling is the model most suited to mixed fleets, and it is the shape that most often lowers enterprise IoT connectivity costs when device usage varies widely across the estate. It is why our own plan structures are built around a shared pool with dynamic allocation rather than rigid per-device caps. You can see how those structures shape an IoT connectivity budget on the plans and pricing page.
There is a fourth shape worth naming because it appears in every comparison the moment cost comes up.
Flat-rate or lifetime pricing charges a single upfront fee covering a fixed data volume over a fixed number of years, with no recurring bill until that ceiling is reached. 1NCE built its position on this model. The trade is predictability against a hard ceiling: below the included volume it is genuinely cheap and administratively simple, and above it you are buying top-ups at whatever the prevailing rate is. It fits low-data sensors with a well-understood duty cycle and fits poorly anywhere payload size might grow. In an IoT connectivity budget it behaves more like a hardware line with a usage ceiling attached than like a monthly service.
| Model | Bills on | Best fit | Main budget risk |
|---|---|---|---|
| Per-MB (pay as you use) | Actual data plus a per-SIM fee | Low or unpredictable use | Uncapped tail devices |
| Per-device-per-month | Fixed allowance per SIM | Uniform, predictable fleets | Over-buying headroom, overage rates |
| Pooled or dynamic | One shared volume per account | Mixed fleets with a wide usage spread | Pool sized on optimistic averages |
| Flat-rate or lifetime | One upfront fee, fixed ceiling | Low-data sensors, long life, stable payload | Cost after the ceiling is reached |
What Per-Megabyte Pricing Actually Depends On
The most common request we see is for a like-for-like per-MB comparison between named providers. It is a reasonable thing to want, and it is close to impossible to get honestly, because the published number is a different number at each vendor. It is also where most IoT connectivity budgeting stalls.
Hologram, emnify, Soracom, KORE, Onomondo, Wireless Logic, Aeris, Verizon and AT&T all publish or quote per-MB figures at some point in the sales process. Those figures are not comparable out of the box, for five specific reasons:
Geography is baked in. A rate quoted for the United States and a rate quoted as a global roaming rate are different products. The same SIM in Mexico, Brazil or Turkey can bill at a multiple of its home-country rate, or be unavailable on a permanent-roaming basis at all.
The SIM fee is sometimes in and sometimes out. A low per-MB rate paired with a monthly per-SIM access fee can cost more than a higher per-MB rate with no access fee, at every volume below a crossover point you have to calculate yourself. Two providers can quote what looks like identical IoT connectivity pricing and bill you very differently for the same fleet.
Rounding rules differ. Some platforms bill in 1 KB increments, some round every session up to 10 KB or 100 KB. On a device sending 200-byte payloads every five minutes, a 100 KB rounding rule can multiply billed volume by more than an order of magnitude against actual bytes sent. This is the single most under-examined variable in IoT connectivity costs.
Volume tiers are not published. The public rate is the top of the tier ladder. What matters is the rate at your committed volume, which only appears in a quote.
Bundled services differ. Platform access, API calls, SMS, activation, and technical support are inside the rate at some providers and billed separately at others.
The practical answer is to stop comparing published rates and start comparing quotes against one fixed specification. Write down device count, per-device monthly data at the 50th and 95th percentile, countries of operation, expected term, security requirements, SMS volume, and growth curve, then send the identical document to every provider on your list. Any number that comes back without those inputs is a guess, and an IoT connectivity budget built on it will not survive its first quarter.
We know how rarely that happens, because we watch it not happen. In our SEO review call on 10 April 2026, our CEO described losing a shortlist where the buyer had emailed four providers at once: the first thing they ask for is the price, and there is very little time to get into discovery. The buyer was comparing four numbers that had been produced from four different sets of assumptions, and had no way to know it. Nine months earlier, on 15 January 2026, the same pattern played out on an inquiry estimating a need of 7,000 SIMs per year that went nowhere because the buyer was interested in one thing only, which was the price. Volume that large should have commanded a serious negotiation. Instead it collapsed into a single comparison that nobody could win on merit.
If you want the variables behind IoT connectivity costs in more depth, we have written them up separately in a piece on the factors that move IoT SIM pricing. A companion piece on M2M SIM card costs and features works through the same IoT connectivity budget questions for older M2M estates.
Roaming Versus Local Rates: The Cross-Border Cost Trap
The same megabyte costs different amounts on different sides of a border, and the mechanism is worth understanding because it drives a large share of unexpected enterprise IoT connectivity costs.
Roaming pricing applies when a SIM issued by one operator connects through a partner network in another country. The visited network charges your home operator a wholesale rate, and that rate is passed through to you with a margin. Rates vary by country and by partner agreement, so a fleet crossing three borders can see three different effective rates against one plan.
Local breakout routes a device’s traffic out to the internet in the country where it is connected, rather than backhauling it to a home network first. It reduces latency and, depending on the commercial arrangement, can reduce cost by avoiding some of the transit that roaming implies.
Multi-IMSI and eUICC localization go further. A multi-IMSI SIM carries two or more network identities and can switch to the one that behaves as a local subscriber in that market, which means local rates rather than roaming rates and no exposure to permanent-roaming restrictions. Countries including Brazil, Turkey and India restrict permanent roaming, so this is a compliance question as well as a pricing one.
For North American fleets specifically, the cheapest answer for an IoT connectivity budget is often not global at all. A truck that runs Texas to Ontario with occasional Mexico legs may be better served by native single-network data plans on a carrier with strong regional coverage, which trades global flexibility for lower latency and better regional economics. Trafalgar Wireless offers both native single-network plans and multi-network plans as IoT connectivity options for fleet and logistics operators who need cross-border coverage without a separate carrier contract in each country, and which one wins is a route-map question rather than a blanket default.
The budgeting rule: never model a cross-border deployment on a domestic rate. Ask for the rate in every country on your route map, in writing, and ask what happens when a device sits in one of them for more than 90 consecutive days.
The Line Items That Do Not Appear in the Headline Rate
This is where budgets break. Every item below is a real charge at some providers and included at others, which is exactly why the headline rate tells you so little about the true cost of IoT connectivity.
Activation and provisioning fees. A one-time charge per SIM to bring it into service. At scale this is a meaningful capital line: 5,000 SIMs at any per-unit activation fee is a number your finance team will notice. Ask whether activation, suspension and reactivation are charged, because devices that cycle in and out of service can be activated more than once.
Platform and API access. Some providers bill for the connectivity management portal, for API calls above a threshold, or per user seat. Each of those belongs in the IoT connectivity budget as its own line. Our own model does not charge separately for portal access, APIs, activation or suspension, and we made the case for a management layer in a post on running a SIM estate at scale.
Private APN and VPN. A private APN is a dedicated closed network between your devices and your business systems that isolates your traffic from the public internet. A site-to-site IPSec VPN is an encrypted tunnel between your network and the provider’s. Both are commonly recurring charges, sometimes per APN and sometimes per connection, and both belong in the monthly IoT connectivity budget rather than in a security annex. The specifics of what these services cover are set out on our IoT security services page.
Static IP addressing. Devices that need to be reachable inbound, or that terminate a VPN, usually need a fixed private IP. This is typically billed per SIM per month, and across a large estate it becomes a material share of enterprise IoT connectivity costs.
SMS and voice. Even data-first deployments often use SMS for wake-up commands or fallback. It bills separately, per message, and in both directions.
Overage and throttling policy. Overage is what you pay above your allowance. Throttling is what happens to your throughput instead. Neither is inherently worse, but you need to know which one you have bought, because a throttled fleet is an outage that does not appear on an invoice. Trafalgar Wireless does not throttle service usage, which means the trade in our model sits in the data plan rather than in degraded performance for the growth businesses and solution providers we serve.
SLA and uptime guarantees. A service level agreement is a contractual commitment to a level of availability, usually with service credits when it is missed. Credits are rarely large enough to cover the commercial damage of an outage, so read an SLA as a statement of how seriously a provider takes its own network rather than as insurance. Ask for the measurement method, the exclusions, and how a claim is filed.
Decommissioning. Ending a contract can carry a final-invoice period, and moving to a new provider can require re-provisioning or physical SIM replacement if your hardware is not eUICC capable.
How to Build the Budget: Two Worked Models
Here is the arithmetic. Both models turn a deployment description into an IoT connectivity budget, and both use your quoted rates rather than invented ones, because the structure is what transfers and the rates are what you have to go and get. The data volumes are calculated from stated assumptions, which you should replace with measurements from your own pilot.
Model one: a 150-truck regional carrier running North America with cross-border Mexico legs
Start with data volume, not price.
- Reporting interval: one position update every 30 seconds while in motion, one heartbeat every 15 minutes while parked.
- Payload: roughly 250 bytes per position report after protocol overhead on a small message.
- Duty cycle: 10 hours in motion per working day, 22 working days per month.
In motion: 120 reports per hour times 10 hours times 22 days equals 26,400 reports, at 250 bytes each equals about 6.6 MB. Parked and weekend heartbeats add roughly 1 MB. Diagnostic and event traffic, plus TLS session setup on devices that reconnect frequently, realistically doubles the total on many trackers. Call it 15 MB per truck per month as a working figure, and 45 MB for the 95th-percentile truck that reconnects constantly or runs longer days.
Fleet volume: 150 trucks times 15 MB equals 2.25 GB per month at the average. Sized per device at the 95th percentile instead, you would buy 150 times 45 MB, or 6.75 GB. That gap, three times the volume for the same fleet, is the entire argument for pooling, and it is the clearest single lever available in IoT connectivity budgeting for a mixed fleet.
Now layer the recurring lines that make up the monthly IoT connectivity cost:
- Monthly connectivity: (150 SIMs times the per-SIM fee) plus (pool volume times the per-GB rate), or (150 times the per-device bundle price) if you go bundled.
- Cross-border: identify how many of the 150 trucks actually cross into Mexico and for how many days. If 20 trucks run Mexico legs, ask for the Mexico rate specifically and model those 20 lines separately rather than applying a blended rate across all 150.
- Security: if the telematics platform terminates a VPN, add (150 times the static IP charge) plus the monthly VPN or private APN charge.
- Firmware: one 8 MB image per truck per quarter is 1.2 GB per quarter, or 400 MB per month amortized, roughly 18% on top of steady-state volume. Do not leave it out.
Year-one total: (12 times the monthly figure) plus (150 times any activation fee) plus SIM hardware. That is the complete IoT connectivity budget for year one, and every term in it is something a provider can be asked to quote. If you are running video telematics rather than position tracking, every volume number above changes by an order of magnitude and the pooled model becomes close to mandatory. The same IoT connectivity budget structure applies to yard, warehouse and trailer assets, and we have set out that wider picture for logistics and freight operations separately.
Model two: a 5,000-unit embedded sensor deployment
- Reporting interval: one reading per hour, batched into one transmission per day.
- Payload: 24 readings at roughly 80 bytes each, plus session overhead, is about 3 KB per day.
- Monthly volume per device: roughly 90 KB, call it 150 KB with retries and occasional reconnection.
Fleet volume: 5,000 devices times 150 KB equals about 750 MB per month across the entire estate. That is a rounding error in data terms, which means the per-MB rate is almost irrelevant to this IoT connectivity budget. The cost is entirely in the per-SIM monthly fee, the activation fee, and the SIM hardware.
The arithmetic that matters here:
- Monthly: 5,000 times the per-SIM fee, plus a small pooled volume charge.
- One-time: 5,000 times the SIM unit cost, plus 5,000 times any activation fee.
- Device-count drift: budget 5,000 billed lines at 106% of deployed units to cover spares and RMA stock.
- Rounding: with 3 KB daily sessions, a 100 KB session rounding rule would bill this fleet as if each device used 3 MB a month rather than 150 KB. Confirm the rounding rule in writing before you compare anything.
For low-data deployments of this shape, flat-rate lifetime pricing is worth modelling as an alternative, because the included ceiling may cover the device’s entire service life. Run both models and compare total spend over the planned deployment years, not over one month. This is also where budgeting for IoT connectivity intersects with hardware choice: embedded and industrial SIMs cost more per unit and last far longer, which changes the denominator in every per-year calculation you make.
Contract Terms That Change the Real Cost
The commercial terms around the rate frequently matter more than the rate.
Minimum commit. A committed monthly volume or revenue floor buys a better rate and transfers forecasting risk to you. Model what you pay in the months you undershoot.
Term length. Longer terms buy lower rates and cost you the ability to respond when your traffic profile changes. A one-month minimum term, which is what we offer, costs a little in headline rate and buys the option to change your mind. Trafalgar Wireless works with enterprise IoT teams and growth businesses on one-month minimum contracts precisely because early-stage deployments almost never consume what the first model predicted.
One contract, one invoice, one currency. Multi-country deployments assembled from separate carrier relationships generate separate invoices in separate currencies with separate settlement dates, and the finance overhead of reconciling them is a genuine, if unglamorous, cost. Consolidating a multi-country estate behind a single agreement removes that work and makes the IoT connectivity budget legible to the people who have to approve it.
eUICC portability and exit rights. If your SIMs are eUICC capable, you can change operator profiles over the air rather than sending someone to physically swap hardware. If they are not, your switching cost includes a field visit per device. Establish this before you buy hardware, not at renewal, because it is the largest hidden item in a multi-year IoT connectivity budget. Our explainer on connectivity architecture for large deployments covers where these choices bite at scale.
Pre-deployment testing. Ask whether you can validate connectivity on real devices before billing starts. A small free-data allowance per SIM for pre-activation testing, which is how we handle it, removes the cost of discovering an integration problem after the meter is running.
What to Ask a Provider Before You Sign
Send this list, in writing, to every provider you are comparing. The answers are what make an IoT connectivity budget defensible.
On the rate. What is the per-MB or per-device rate at my committed volume, in each country on my route map? Is a per-SIM monthly fee included or additional?
On rounding. What is the billing increment, and is it applied per session or per billing period?
On the plan shape. Is data pooled across the account or allocated per device? What happens to unused volume at month end?
On overage. What is the overage rate, and is there throttling instead of or in addition to overage? Can I set hard caps and alerts per device?
On one-time charges. Is there an activation fee, a reactivation fee, a suspension fee, or a minimum billing period per SIM?
On the platform. Is portal access, API access, and user seating included? Are there API rate limits or call charges?
On security. What are the recurring charges for a private APN, a site-to-site VPN, static IP addressing and IMEI locking, and are they billed per SIM or per APN?
On the SLA. What availability is committed, how is it measured, what are the exclusions, and what is the credit mechanism?
On the term. What is the minimum contract period, the notice period, and the early termination charge?
On the exit. If I leave, can I move my SIMs to another provider’s profile over the air, or does it require hardware replacement?
On support. Who do I contact when a device will not attach at 2am, and is that included?
Frequently Asked Questions
How much does IoT connectivity cost per device per month?
There is no single answer, and any provider that gives you one without asking questions is quoting a fiction. Enterprise IoT connectivity costs are driven by monthly data volume, the countries the device operates in, the total number of devices, the plan shape, and the security and platform services attached to the connection. A low-data sensor sending a few hundred kilobytes a month and a video telematics unit sending several gigabytes are different products with different economics, even on the same network. Size your data volume first using the arithmetic above, then request quotes against that specification.
Is flat-rate lifetime IoT pricing cheaper than per-megabyte pricing?
Below the included data ceiling, usually yes, and it removes recurring billing administration entirely. Above the ceiling you are buying top-ups at prevailing rates and the advantage disappears. The crossover point is where your projected lifetime data use crosses the included volume, so calculate your expected total consumption across the full deployment life and compare it against the ceiling before you decide. Deployments with any prospect of payload growth, added sensors or richer firmware tend to cross it earlier than planned, so model both shapes against your IoT connectivity budget over the full deployment life rather than over a single year.
What is a reasonable data allowance for a fleet tracking device?
For basic position tracking at 30-second intervals during a 10-hour driving day, the arithmetic above lands around 15 MB per vehicle per month, with heavy users near 45 MB. Add firmware updates separately. Devices that stream sensor telemetry, carry driver-facing applications or upload images sit far above that range, and video telematics is a different category again. Measure it on a pilot rather than trusting a spec sheet, because protocol overhead and reconnection behaviour vary more between devices than payload size does.
Does a private APN cost extra?
Usually yes, as a recurring line item rather than a one-time setup. The price is driven by whether the APN is dedicated to you or shared, whether it terminates in a VPN, how many connections sit behind it, and whether you need static IP addressing on those connections. Ask for the charge to be quoted per APN and per connection separately, so you can see how it scales as the fleet grows and where it lands in the IoT connectivity budget.
How do I compare providers on activation speed?
Ask two concrete questions: how long from order to SIMs arriving with active profiles, and how long from a SIM being activated in the portal to the device attaching to a network. The second matters operationally, because it determines whether a technician can commission a device on site or has to return. Providers offering in-factory profile provisioning can ship devices that connect on first power-up, which removes an entire commissioning step from the deployment cost.
Key Takeaways
Model volume before price. The data arithmetic determines which pricing model wins. Rates negotiated against the wrong plan shape save you very little.
Pool where usage varies. Sizing every device for its worst month can triple the volume you buy compared with a shared pool serving the same fleet.
Get the rounding rule in writing. Session rounding can multiply billed volume against actual bytes sent on small-payload devices, and it is almost never mentioned in a headline rate.
Price the line items outside the rate. Activation, platform access, private APN, VPN, static IP, SMS and overage decide the real total more often than the per-MB figure does.
Quote the route map, not the country. Cross-border deployments need a rate per country and a clear answer on permanent-roaming restrictions before the budget is signed off.
Treat the exit as part of the entry. Term length, termination charges and eUICC portability determine what your renewal negotiation looks like in two years.
Send one specification to everyone. Identical inputs are the only way to produce comparable quotes, and comparable quotes are the only defensible basis for IoT connectivity budgeting.
Closing Thought
The best evidence for all of this arrives in our inbox. On 11 May 2026, in a review call, our CEO read out a genuine inbound inquiry word for word: 30 SIM cards on a private APN, OpenVPN across two or three of them for a secure private connection, an average of under 300 megabytes per SIM, can you provide the service and what will the cost be. His comment afterwards was that there was not much information there, and he was right. That message is more specific than most, and it still omits the countries, the contract term, the growth curve, whether SMS is needed, whether static IPs are required, and what happens when the estate goes from 30 SIMs to 300. Any number sent back against it would have been a guess dressed up as a quote.
The fix is not a better rate card. It is a specification good enough that every provider has to price the same deployment, and a budget model that carries the recurring lines, the one-time lines and the exit cost in the same spreadsheet. Build that once and you can compare anyone.
If you want that conversation to start with your deployment rather than with a number, Trafalgar Wireless works through IoT connectivity budgeting with enterprise IoT teams and growth businesses planning multi-country deployments, providing multi-network IoT data plans across 500+ networks in 180+ countries under one contract, one invoice and one currency.
