Request a firmware review
Menu

LTE-M or NB-IoT in 2026: choosing cellular for US deployments

Reader question: For a cellular device deployed in the US, and perhaps abroad, should we design for LTE-M, NB-IoT or both, given current carrier support?

Micrograph-style illustration of two winged seeds side by side, a broad-winged ash seed and a narrow-winged birch seed, with the birch seed tinted amber.

For a device deployed in the US, the LTE-M vs NB-IoT answer is to design for LTE-M first. As of November 2026, AT&T’s own LPWA page says LTE-M is available on its US network and that “NB-IoT is available outside the United States”, while Verizon’s page lists both LTE-M and NB-IoT as supported on its 4G LTE network. If the device moves, needs low latency or must work on both AT&T and Verizon, LTE-M is the safer base; if you also ship abroad, or have stationary devices where a carrier offers NB-IoT, choose a module that supports both and keep your application and update traffic within what NB-IoT can carry.

The rest of this article explains why, starting from what the carriers say today and then comparing LTE-M vs NB-IoT on what a firmware team actually feels: handover, data rate, power-saving timers, coverage and the time it takes to push a firmware update. Network availability changes, so every carrier statement here carries the date it was checked.

Fumbar is a Denver firm that writes embedded firmware and over-the-air update systems for hardware startups, including cellular devices. Nothing below is a recommendation of a specific carrier or module vendor; check every network fact with the carrier you will actually use.

The short answer for a US-first device

Pick the technology your carrier supports where your devices will sit, then pick the one whose limits your firmware can live with. For most US startups that leads to LTE-M, with NB-IoT as an option for export or for stationary devices on a carrier that supports it.

A few definitions first. LTE-M is the name carriers use for LTE category M devices; Verizon’s page writes it as “LTE-M (Cat-M1)”, and 3GPP’s specifications call these devices bandwidth-reduced low-complexity (BL) UEs, in categories M1 and M2. NB-IoT (narrowband IoT) uses UE categories NB1 and NB2. Both are low-power wide-area (LPWA) cellular technologies that run on carriers’ LTE networks.

Your situation Design for Why
US only, device moves (trackers, vehicles, wearables) LTE-M NB-IoT does not support handover in connected mode
US only, stationary, must work with AT&T LTE-M AT&T lists NB-IoT as available only outside the US
US only, stationary, carrier confirmed to support NB-IoT where you deploy Either; dual-mode module preferred Keeps the option to move carriers
US first, export planned Dual-mode module, LTE-M first in the US Roaming and local network support differ by country
Needs voice or emergency calling LTE-M 3GPP lists real-time services and emergency call as not supported for NB-IoT
Large firmware images, frequent updates LTE-M Larger transport blocks and wider channel mean shorter transfers

The table is a starting point, not a substitute for the carrier’s own confirmation. The chapters below explain each row.

Example (fictional). Pinewater Telemetry, a Denver startup, builds a water-level sensor for stock tanks on Colorado ranches. Each unit is bolted to a tank, reports six readings a day, runs on lithium primary cells and must last several seasons. The founders plan to launch in Colorado and Wyoming and later sell into Canada. Stationary and delay-tolerant, the sensor fits the profile Verizon describes for NB-IoT, but the team wants the option to use AT&T, whose page lists NB-IoT only outside the US. They choose a module that supports both, set LTE-M as the first choice in the US, keep the technology order configurable over the air, and test both at five real tanks before freezing the board.

What US carriers say, as of November 2026

The carriers’ own pages are the only reliable source for availability, and they disagree. AT&T supports LTE-M in the US and NB-IoT only abroad; Verizon lists both; T-Mobile’s pages could not be read for this article.

AT&T’s LPWA page, checked November 22, 2026, says: “LTE-M is currently available on the AT&T network in the US and roaming carriers in select other countries.” Under “NB-IoT Worldwide Availability” it says: “NB-IoT is available outside the United States.” The same page’s plan terms say “Must use AT&T approved IoT LTE-M device”, a reminder that carrier device approval is part of the schedule. For export planning, the page’s “LTE-M Worldwide Availability” list names more than forty countries and territories, including Canada, Mexico and Puerto Rico, which helps a US-first product that will cross a border later.

Verizon’s LTE-M and NB-IoT page, checked the same day, says the two technologies “are both fully supported by our 4G LTE network.” Its FAQ positions NB-IoT for low-power, delay-tolerant, stationary sensors that transmit rarely and “Do not require connected mobility (handover) or voice support”, and describes LTE-M (Cat-M1) as designed for devices “that transmit small to medium amounts of data over wide ranges.” The page also lists Non-IP Data Delivery (NIDD) support, a 3GPP feature defined in TS 23.682.

T-Mobile’s business IoT pages and the GSMA’s Mobile IoT deployment map both refused automated requests (HTTP 403) when this article was researched, so neither is quoted here. If T-Mobile is on your list, get its current LTE-M and NB-IoT status in writing from its IoT sales team.

CarrierLTE-M in the USNB-IoT in the USWhat the page says
AT&TYesNo"LTE-M is currently available on the AT&T network in the US"; "NB-IoT is available outside the United States"
VerizonYesYesLTE-M and NB-IoT "are both fully supported by our 4G LTE network"
T-MobileNot verifiedNot verifiedBusiness IoT pages returned HTTP 403 to automated requests; ask the carrier
Figure 1. LTE-M and NB-IoT status as stated on carrier pages, checked November 22, 2026. Sources: business.att.com/products/lpwa and Verizon's LTE-M and NB-IoT page. Availability changes; confirm with the carrier before you commit to a design.

Two cautions about reading older material. News coverage from the launch years describes plans and pilots that may no longer hold, so rely on the carrier’s current page and a written answer from its IoT team. And “available on the network” does not mean “available at your site”; coverage maps and a field test at real install locations decide that.

How to check your carrier in an afternoon

  1. Read the carrier’s current IoT network page and save a dated copy for your design record.
  2. Ask the carrier’s IoT team, in writing, which of LTE-M and NB-IoT it supports at your deployment sites, and whether it supports power saving mode and extended idle mode DRX with the timer values you plan to request.
  3. Ask what device approval the carrier requires for your module and product, and how long it takes.
  4. Put a development kit with your intended module at two or three real install sites and log attach success, signal quality and the timers the network actually grants.

How long will these networks last?

None of the carrier pages read for this article gives an end date for LTE-M or NB-IoT service, and this article does not guess one. On the standards side, both remain in the current 3GPP specifications: the Release 19 versions of TS 36.300 and TS 36.306 published by ETSI in 2026 still define BL UEs, enhanced coverage and NB-IoT. TS 36.300 also notes that “interworking with NR is not supported by BL UE”, so an LTE-M device does not hand over to or measure 5G NR cells. Verizon’s page describes its LTE-M and NB-IoT offerings as “5G compatible”; ask what that means for your module in writing.

For a product with a long support period, ask each carrier for its written network roadmap for the technology you choose, and design so that a module change is a board revision and a driver, not a rewrite. Portable firmware architecture covers how to keep the modem behind an interface that survives a part change.

Two technologies from the same 3GPP release

LTE-M and NB-IoT were standardized together. 3GPP announced in June 2016 that it had “completed the standardization of NB-IOT, the new narrowband radio technology developed for the Internet-of-Things”, implemented in Release 13 (LTE Advanced Pro). The same announcement says the portfolio of new technologies operators could use “includes NB-IOT, eMTC and EC-GSM-IoT”. LTE-M is commonly written as LTE-M (3GPP eMTC), the second of those.

The core difference is bandwidth. 3GPP TS 36.300 (ETSI version 19.2.0, August 2026) describes a BL UE, the LTE-M device class, as operating “with a limited channel bandwidth of 6 PRBs (corresponding to the maximum channel bandwidth available in a 1.4 MHz LTE system).” For NB-IoT, it says the “full carrier bandwidth is 180 kHz”, with subcarrier spacing of 3.75 kHz or 15 kHz, and calls its physical layer “optimized for very low power consumption.”

Both have been extended since Release 13. Category M2 devices can use a 24 PRB channel for higher data rates, and category NB2 raises NB-IoT’s transport block sizes. The current specification also covers both over non-terrestrial (satellite) networks: TS 36.300 says E-UTRAN “supports radio access over non-terrestrial networks for BL UEs, UEs in enhanced coverage and NB-IoT UEs.” Whether a given carrier offers satellite service to your module is, again, a question for the carrier.

The table below puts the differences that matter to firmware side by side, with the source for each row.

Property LTE-M (BL UE, Cat M1 and M2) NB-IoT (Cat NB1 and NB2) Source
Channel or carrier bandwidth 6 PRBs (1.4 MHz); M2 optionally 24 PRBs for data 180 kHz TS 36.300
Handover in connected mode Supported Not supported TS 36.300 (NB-IoT exclusions); Verizon FAQ
Real-time services and emergency call Supported Not supported TS 36.300 (NB-IoT exclusions); Verizon FAQ
Largest downlink transport block 1000 bits (M1), 4008 bits (M2) 680 bits (NB1), 2536 bits (NB2) TS 36.306
Power saving mode and eDRX Yes, network grants the values Yes, network grants the values TS 23.682
Early data transmission during random access Yes Yes TS 36.300
AT&T, US, November 2026 Available Listed as outside the US only AT&T LPWA page
Verizon, US, November 2026 Supported Supported Verizon LTE-M and NB-IoT page

What NB-IoT leaves out is the part that matters most to firmware. TS 36.300 lists functions “not supported for NB-IoT”, including “inter-RAT mobility, handover, measurement reports”, “real-time services” and “emergency call.” Those exclusions are listed for NB-IoT only; LTE-M devices are not subject to them, which is why Verizon steers devices that need handover or voice to LTE-M.

Mobility: the difference firmware feels first

NB-IoT does not hand a connection over from one cell to the next. TS 36.300 states it plainly: “For NB-IoT: Handover, measurement reports and inter-RAT mobility are not supported.” A device that moves while connected has to drop to idle and reselect a cell, then reconnect and resume.

For a stationary meter, that rarely matters. For anything that moves while it talks, it shapes the whole communication design:

  • Asset trackers and vehicles. A transfer that spans a cell boundary breaks and must restart or resume. Short messages usually survive; long transfers, such as a firmware image, need resumable downloads at the application layer either way, and need them more on NB-IoT.
  • Wearables and patient devices. Someone walking or driving crosses cells. Verizon’s own FAQ steers devices that need “connected mobility (handover)” away from NB-IoT.
  • Devices that must call for help. Emergency call and real-time services are on the NB-IoT not-supported list, so voice-capable alarms and personal emergency devices belong on LTE-M.

Even on LTE-M, mobility costs power: the device measures neighbor cells and the network may move it between them. A tracker that reports every few minutes and sleeps in between spends most of its life in idle mode, where both technologies reselect cells on their own.

For trackers specifically, decide early how a device reports position while moving. Batching several fixes into one message sent from a stationary moment, or at a fixed interval, spends less airtime than streaming, and it lets the modem stay idle while the vehicle crosses cells. If the product promises near-real-time location, that promise points to LTE-M.

The firmware lesson is the same on both: treat a connection as something that can end mid-transfer at any moment. Design messages to be idempotent (safe to receive twice), give every upload a sequence number, and make firmware downloads resumable from a byte offset.

Data rate and latency: what the specifications allow

LTE-M can move far more data per scheduling interval than NB-IoT, and that difference shows up in latency, transfer time and energy per byte. The clearest primary numbers are the transport block limits in 3GPP TS 36.306, which define the most data a device category can receive or send in one transmission time interval.

UE category Largest downlink transport block (bits) Largest uplink transport block (bits) Layer 2 buffer (bytes) Notes from TS 36.306 and TS 36.300
M1 (LTE-M) 1000, or 1736 with an optional feature 1000, or 2984 with an optional feature 20,000 or 40,000 6 PRB channel (1.4 MHz)
M2 (LTE-M) 4008 6968 100,000 Up to 24 PRBs; a UE indicating M2 also indicates M1
NB1 (NB-IoT) 680 1000 4,000 180 kHz carrier
NB2 (NB-IoT) 2536, or 4968 with 16QAM support 2536 8,000 or 12,000 A UE indicating NB2 also indicates NB1

Transport block sizes are not throughput. Real throughput depends on how often the network schedules the device, how many repetitions it needs in poor coverage, half-duplex operation and protocol overhead. NB-IoT’s 180 kHz carrier and smaller transport blocks also mean each message occupies the radio for longer, which adds latency. Module datasheets quote peak rates measured under ideal conditions; treat them as upper bounds and measure your own.

For a sensor that sends a few hundred bytes an hour, both technologies are fast enough, and the difference that matters is energy per report and cost. For anything that occasionally moves a lot of data, such as logs, images or firmware, LTE-M’s larger blocks and wider channel shorten the time the radio is on, which is often the biggest single draw on the battery.

Latency matters for interactive features too. A command from your cloud to a device that is asleep in a power-saving mode waits until the device next listens, on either technology. Waking more often costs energy, so the latency you promise customers and the battery life you promise them are set by the same timers.

When neither fits: LTE Category 1bis

Some products outgrow both. A camera trap, a device that uploads audio, or a gateway that relays a whole site’s data may need more than LTE-M carries comfortably. LTE Category 1bis sits between LTE-M and full LTE. In TS 36.306 its largest downlink transport block is 10296 bits and its largest uplink block 5160 bits, its layer 2 buffer is 150,000 bytes, and a device indicating it “is only required to support 1Rx antenna”, which keeps the antenna design closer to an LPWA product.

The trade is power and coverage. Category 1bis devices do not use the narrower channels of LTE-M and NB-IoT, and Verizon’s page notes that NB-IoT and LTE-M “offer greater coverage than you’d get with LTE Cat-1.” If your duty cycle is light and your messages are small, Category 1bis spends energy on capacity you do not use; if your data is heavy, it can finish faster and sleep sooner. Measure both on your actual traffic before deciding.

Power: PSM and eDRX work on both, and the network decides

Both LTE-M and NB-IoT devices can use power saving mode (PSM) and extended idle mode DRX (eDRX), defined in 3GPP TS 23.682. The device requests the timers; the network grants what it chooses. Your firmware must read back what was granted and plan around it.

PSM. TS 23.682 says PSM “is similar to power-off, but the UE remains registered with the network and there is no need to re-attach or re-establish PDN connections.” A device in PSM “is not immediately reachable for mobile terminating services”; it can be reached while connected and for an Active Time afterwards. The specification calls PSM “intended for UEs that are expecting only infrequent mobile originating and terminating services and that can accept a corresponding latency in the mobile terminating communication.”

eDRX. Extended idle mode DRX stretches the interval between the moments an idle device listens for paging. The device requests eDRX parameters during attach and tracking area updates, and the network “may also provide different values of the extended idle mode DRX parameters than what was requested by the UE.” TS 23.682 adds that a UE “should request for extended idle mode DRX only when all expected mobile terminating communication is tolerant to delay.”

Three design rules follow:

  1. Read the granted values. The network may shorten or lengthen what you asked for, or decline. Log the granted Active Time, periodic update timer and eDRX cycle with every report, so the fleet data shows what each network actually does.
  2. Let the device poll. TS 23.682 notes that if a server “has periodic downlink data, it is more efficient when the UE initiates communication” to poll for it. For commands and update offers, have the device ask when it wakes instead of the server pushing into silence.
  3. Measure, then estimate. Datasheet sleep currents assume ideal timers and good coverage. Measure average current on real hardware at real sites with the timers the network grants; measuring sleep current covers the method and why battery life falls short of the estimate covers what estimates miss.

If you plan to wake devices by SMS or a carrier device trigger, size the Active Time with care. TS 23.682 gives a minimum recommended Active Time for SMS delivery of “2 DRX cycles plus 10 seconds”, and notes that a network may grant a shorter Active Time, even 0 seconds. A device that asks for a very short Active Time to save energy can miss the very message meant to wake it, so test the trigger path on the real network with the granted values.

The battery life calculator turns measured currents and wake schedules into an estimate, and it is a better guide than any network-level battery claim. Both carrier pages quote battery lives of up to 10 years or more; those figures depend on the application, as Verizon’s own wording says.

Coverage, and where NB-IoT still helps

Both technologies are aimed at places ordinary phones struggle, such as basements and meter pits, and carriers market them that way: AT&T says LTE-M “is available in challenging environments such as underground areas and deep inside buildings”, and Verizon says NB-IoT and LTE-M “offer greater coverage than you’d get with LTE Cat-1.”

The cost of deep coverage is time on air. In enhanced coverage the network repeats transmissions, so each message takes longer and uses more energy, and a firmware download that takes seconds in good coverage can take far longer in a meter pit. Plan for the worst install location you intend to support, not for the bench.

For LTE-M, TS 36.300 defines two enhanced coverage modes, mode A and mode B, and says “the support of enhanced coverage mode A is mandatory for a BL UE.” The network publishes, per coverage enhancement level, how many times a device repeats its random access attempts, and devices at the same level share the same resources. If your module reports the coverage enhancement level it is using, log it with every report so your fleet data shows which installs are expensive in time and energy, and which need an external antenna.

NB-IoT still has a place in a US-first plan:

  • Stationary devices on a carrier that supports it. A water meter or a tank sensor that never moves and reports a few times a day is the profile Verizon describes for NB-IoT.
  • Export. AT&T’s own page lists NB-IoT as available outside the United States, so a product sold abroad may meet it; check each market’s carriers.
  • Fallback. A dual-mode module can use NB-IoT where LTE-M is weak or absent, if the carrier and SIM plan allow both.

Coverage is local. Before you choose, put a development kit at a handful of real sites and log signal quality, attach time and the coverage enhancement level the module reports.

Small messages: early data, NIDD and the IP stack

Both technologies include ways to send small amounts of data without the full cost of a normal connection, and they matter more to battery life than the headline data rate. Whether you can use them depends on the carrier, so treat them as options to confirm, not assumptions.

Early data transmission. TS 36.300 describes mobile-originated early data transmission (MO-EDT), which “allows one uplink data transmission optionally followed by one downlink data transmission” during the random access procedure. It applies when the uplink data is no larger than a transport block size the cell advertises, and the random access form “is only applicable to BL UEs, UEs in enhanced coverage and NB-IoT UEs.” For a sensor that sends one small reading and goes back to sleep, that removes most of the connection setup. Ask the carrier whether it is enabled, and ask the module vendor whether the firmware can request it.

Non-IP Data Delivery. TS 23.682 defines NIDD for communication “where the data used for the communication is considered unstructured from the EPS standpoint”, delivered to your server either through the carrier’s service capability exposure function (SCEF) or a point-to-point tunnel. Verizon lists NIDD support on its LTE-M and NB-IoT page. Dropping the IP stack saves code space and bytes on the air, but it moves work to your firmware: there is no TLS on that path, so authenticate and encrypt payloads at the application layer, and confirm with the carrier how you would deliver a firmware image of hundreds of kilobytes.

IP with TLS. The familiar path costs more per message, mainly in handshakes, but works with any server and any carrier. Session resumption, a long-lived session across sleep periods, and compact formats such as CBOR keep the overhead down. Whatever path you choose, design the message format so that a report and its acknowledgment each fit in as few transmissions as possible; on NB-IoT, every extra round trip in deep coverage is seconds of radio time.

Firmware updates over LTE-M and NB-IoT

Every cellular device needs a way to fix its firmware in the field, and the cellular link sets the time, energy and data cost of each update. On LTE-M, a full image is usually practical. On NB-IoT, especially in deep coverage, plan for smaller images, resumable transfers and patience.

The arithmetic is simple. A 412 KB image is 3,375,104 bits; with 8% protocol overhead it is about 3.65 million bits on the air. Figure 2 shows how long that takes at four throughputs. The throughputs are assumptions chosen to illustrate the range, not measurements of any network or module.

Time to transfer a 412 KB image (assumed throughputs) 300 kbit/s LTE-M, good signal about 12 seconds 50 kbit/s LTE-M, weak signal about 73 seconds 20 kbit/s NB-IoT, good signal about 3 minutes 2 kbit/s NB-IoT, deep coverage about 30 minutes
Figure 2. Transfer time for a 412 KB image plus 8% overhead at four assumed application throughputs, bars drawn to scale. Hypothetical throughputs for illustration only, not measurements; measure your module on your carrier at real install sites.

Data cost scales with fleet size. The default example in the over-the-air data cost calculator assumes a hypothetical fleet of 5,000 devices, a data price of $0.40 per MB, 8% overhead and 5% of devices downloading twice. One rollout of a 412 KB full image costs about $913, and a delta image at 30% of that size about $274. Delta firmware updates explains when that saving is worth the added work on the device and in the build.

Energy follows time on air. Measure the module’s average current once during a real download at a weak-signal site, multiply by the transfer time, and add that to your battery budget for every release you expect to ship in the product’s life. On a primary-cell device, a few long updates in deep coverage can cost more than months of normal reporting.

Four firmware rules matter more on cellular than on any other link:

  1. Resume, never restart. Store the image in the spare slot as it arrives and resume from the last verified offset after a dropped connection or a power-saving sleep.
  2. Stay awake on purpose. Request a short power-saving pause or stay connected for the duration of a download, then restore the normal timers; otherwise the device may sleep halfway through.
  3. Check before you download. Confirm the image is meant for this hardware revision and is newer than what is installed before spending airtime on it.
  4. Roll out in waves. A bad release on a cellular fleet costs data twice, once to send it and once to fix it. Staged firmware rollouts covers the halt rules, and what it takes to ship firmware you can safely update covers the rest of the update chain.

Example (fictional). Loamwick’s 40 LW-1 soil probes talk Bluetooth LE to 4 solar-powered gateways, and only the gateways have LTE-M modules. For the v1.1.0 release, each gateway downloads the 412 KB image once over LTE-M and passes it to its probes over Bluetooth LE. With the data cost calculator’s default overhead and retry assumptions, one download is 0.456 MB, so the rollout uses about 1.8 MB of cellular data, against about 18 MB if all 40 probes had their own cellular link. Putting the cellular radio only where there is solar power also keeps the probes’ battery budget free of cellular transmit peaks.

Choosing a module and planning for export

Most startups should buy a certified cellular module rather than design the radio themselves, and for a US-first product with export plans, a module that supports both LTE-M and NB-IoT keeps options open at little firmware cost. Dual-mode modules exist; Digi’s documentation, for example, is titled “Digi XBee 3 Cellular LTE-M/NB-IoT Global Smart Modem User Guide.”

Ask each candidate module vendor and carrier these questions before committing:

Question Why it matters
Which LTE bands does the module support, and which does each target carrier use? A module without your carrier’s bands cannot attach, whatever the technology
Is the module approved, or approvable, on each carrier you need? AT&T’s plan terms require an “AT&T approved IoT LTE-M device”; approval takes time
Which UE categories does it support (M1, M2, NB1, NB2)? Sets the transport block limits in the table above
Does it support PSM and eDRX, and how does it report the granted timers? Your firmware needs the granted values, not only the requested ones
How is the module’s own firmware updated in the field, and who signs it? The modem is a second processor in your product with its own update path
Which roaming partners does the SIM plan include in your export markets? AT&T lists LTE-M roaming carriers in “select other countries”

Plan the SIM with the same care. Decide early whether the SIM is a socket, a soldered chip or embedded in the module, who owns the carrier account, and how a device in the field would move to another carrier if prices or coverage change. Those answers shape the board, the factory process and the support plan, and they are hard to change once units ship.

On the firmware side, keep the radio choice a setting, not a constant. Store the preferred technology order (LTE-M first, NB-IoT second, or the reverse for a given market) and band list in configuration that can change over the air, and report which technology each device is using. If a carrier changes its network, you then change a setting in a release, not the hardware.

Keep the modem behind a narrow interface in your code: connect, send, receive, sleep, report status. Vendor AT command sets differ, and a later module change for supply or cost reasons should not ripple through the application. Radio certification and carrier approval also need specific test modes from your firmware; RF certification test firmware covers what labs ask for.

A check-your-carrier and design checklist

Before you freeze the design, work through this list and keep the answers, with dates, in your design record. Carrier facts are true as of the day you check them, so plan to re-check before each production run.

Network

  • Carrier IoT page read and saved with the date; LTE-M and NB-IoT status confirmed in writing for your deployment areas.
  • Field test at real install sites: attach success, signal quality, coverage enhancement level and granted power-saving timers logged.
  • Carrier device approval requirements and timeline known for your module and product.

Firmware

  • Connections treated as interruptible: idempotent messages, sequence numbers, resumable downloads.
  • Granted PSM Active Time, periodic update timer and eDRX cycle read back and reported with every message.
  • Device polls for commands and update offers when it wakes, instead of relying on downlink pushes.
  • Radio technology order and bands configurable over the air; technology in use reported to the cloud.
  • Coverage enhancement level and signal quality logged with every report.
  • Modem behind a narrow driver interface, so a module change stays inside one layer.

Updates and power

  • Update image size and transfer time checked against your worst install site, with a delta plan if needed.
  • Staged rollout with halt rules before the first fleet-wide release.
  • Average current measured on real hardware at real sites, and the battery estimate rebuilt from the measurements.

If you are building a cellular product and want a second set of eyes on the radio, power and update design, connectivity describes the work we do on cellular and other links, and industrial IoT covers sensors, gateways and retrofit monitoring. Our engagements start with a firmware review: we read the schematics, the current code and the update plan, and deliver a written plan with the risks and the order of work. To start one, request a firmware review.

Facing this decision on a real device?

If this article describes a choice in front of you, a firmware review puts it in writing for your board, your code and your update plan.

Request a firmware review

Or email sales@fumbar.com, or call (720) 675-4202.