Ask an operations manager what they wanted from IoT five years ago and you’ll get the same answer every time: tell me where my stuff is. That question is basically solved. Tag on a pallet, gateway on the ceiling, dot on a map. If location is still your headline feature in 2026, you’re selling a problem the market already closed.
The question that’s actually costing people money now is different, and it’s a lot less photogenic: not did the pallet reach the cold room, but what temperature was it sitting at for the four hours that trailer sat on the dock. Location tells you where a thing is. It says nothing about whether that thing is still any good.
That gap is where a quiet category of hardware has been growing — sensor beacons and data loggers. Not flashy, not on the keynote stage, and increasingly the reason people are buying in the first place.
Location got solved. Condition is still open for business
Worth being straight about how the positioning conversation ended. BLE angle-of-arrival systems now quote around 50-centimetre accuracy in real deployments. UWB does 10 to 30 centimetres reliably and got cheaper. Bluetooth 6.0 added Channel Sounding, and the Bluetooth SIG set a 0.5-metre accuracy requirement for it.
All of that is good news, and none of it answers the question a quality manager asks at 7am when a shipment gets rejected. “Where was it” is a solved problem. “What happened to it” is not.
A cold room, a vaccine fridge, a greenhouse, a clean room, a reefer container in a Rotterdam yard — the assets in all of those places move rarely. Their location is boring and known. Their condition is the entire business risk.
What a “sensor beacon” actually is
There are two kinds of Bluetooth beacon, and conflating them is how projects go sideways.
The first kind broadcasts an identifier — an iBeacon UUID with Major and Minor, or an Eddystone namespace and instance ID. It’s a signpost. Your software reads the signal strength and works out that a phone or a gateway is near signpost number 47.
The second kind broadcasts measurements. Actual numbers, from sensors on the board: temperature, humidity, acceleration, whatever else is mounted. The protocol underneath doesn’t have to change much — it just carries sensor readings instead of only an ID.
KKM’s devices do both at once. All their sensor beacons broadcast a KSensor frame by default (battery level, temperature, humidity, acceleration) while still carrying optional iBeacon and Eddystone UID/URL/TLM frames that are switched off out of the box. That design choice matters more than it sounds: the same tag can feed a temperature monitoring system and a proximity or location system, without buying two devices or maintaining two inventories.
The 2026 numbers that explain the shift
The macro picture is not subtle. Industry analysts put the overall IoT market at roughly $1.3 trillion by 2026, with the smart-city segment alone climbing from $130.6 billion in 2021 to $312.2 billion in 2026 (IoT Insider).
The number that matters most for sensor hardware is the predictive maintenance market, which has grown from $1.5 billion in 2016 to a projected $28 billion by 2026. The reported results from mature deployments are the part worth reading twice: maintenance cost reductions of 25-30%, asset life extension of 20-25%, and downtime reductions of up to 50% (IoT Insider).
None of those outcomes come from knowing where a machine is. They come from knowing what it’s been doing — vibration patterns, temperature drift, door-open counts, the slow curve that tells you a compressor is failing three weeks before it fails.
The Bluetooth hardware market reflects the same pull. The Bluetooth IoT chipset market alone reached $7.2 billion in 2024 and keeps growing alongside wearables and asset tracking (The Network Installers). And on the long-range side, LoRaWAN passed 125 million deployed end devices worldwide at the end of 2025, growing at roughly 25% a year (LoRa Alliance 2025 End of Year Report).
Regulation is doing the selling now
Here’s the part a lot of hardware buyers haven’t internalised yet. The EU’s Cyber Resilience Act isn’t only about security — it’s about evidence.
The compliance calendar is landing right now. Member states had to appoint conformity assessment bodies by 11 June 2026. Incident and vulnerability reporting obligations take effect on 11 September 2026, requiring manufacturers to report actively exploited vulnerabilities within 24 hours. Full compliance lands on 11 December 2027. Penalties run up to €15 million or 2.5% of worldwide annual turnover (IoT Insider).
Stack that against cold chain and pharmaceutical quality regimes — GDP, GSP, GMP — which have always demanded traceable records and have always been satisfied with paper. Put the two together and the ask is identical: if you can’t produce a timestamped, retrievable record, you’re exposed.
There’s a real difference between a dashboard screenshot and a device that holds 60,000 logged data points on board and exports them as PDF or Excel. One of those survives an audit. The other is a story you tell.
The hardware, without the brochure voice
KKM’s sensor line is deliberately unglamorous and reasonably complete. Five devices cover most of what a condition-monitoring project actually needs.
S23 temperature logger — the cheap one you stick on things. IP65, 50.5 × 29.6 × 7 mm, 10.1 grams, CR2032 coin cell good for about 12 months at the default one-second broadcast interval. It measures -20°C to 60°C with ±0.4°C accuracy between 0 and 60°C, holds 60,000 records, and reads 7,000 of them in about 15 seconds through KGateway. Mounts with double-sided tape. If you’re shipping pallets out and want proof of what happened en route, this is the device you tape in the box and forget about.
K4PT BLE temperature logger — the reusable one. IP67, CR2477 replaceable cell rated for more than 48 months, ±0.5°C across -40°C to 70°C, 60,000 logging points. The interesting spec is range: up to 600 metres to a KKM gateway in PHY Coded mode, which is what makes it usable across a yard or a refrigerated dock rather than just inside one room.
S6e temperature and humidity sensor — the fixed-asset one. IP67 to IP68, 8+ years on a 2600 mAh ER14505 lithium cell, ±0.4°C, 60,000+ offline T&H records, 300 metre range. The probe variants are what make it flexible: the S6et1 and S6et2 push measurement range out to 280°C and 200°C respectively on cabled probes, which is food processing and industrial territory rather than cold storage.
LAS6 LoRaWAN temperature and humidity sensor — for the places you can’t run network cable or reach with Bluetooth. IP67, Class A on EU868 and US915, NFC for fast bulk setup, caches readings when the network drops and uploads them on reconnect.
Gateways, because none of this works without them. The KG02 is the indoor workhorse — slim, PoE or 5V DC, CE/FCC certified, hardware watchdog for 24-hour uptime, scanning 200+ beacons per second. The KG01 is the outdoor one: IP54, Bluetooth 5.0 Coded PHY with up to 900 metres of range in open space, 300+ beacons per second, GPS and LTE optional, and local storage so packets survive an outage. Both speak MQTT and HTTPS with TLS, so data lands on your server or your cloud — AWS, Azure, on-prem, wherever.
Where all of this gets tricky
Nobody’s datasheet covers this part, so here it is plainly.
Broadcast interval is a battery budget. That 12-month S23 figure assumes a one-second interval. Coded PHY long range and higher TX power buy you distance and cost you battery. Work out your actual duty cycle before you commit to a device count.
Logging is not alerting. A logger that stores 60,000 points is a recorder. If you need to know about an excursion within seconds, you need a gateway in range continuously — not someone walking past with a phone app. Pick which problem you’re solving.
±0.4°C is a sensor spec, not a system spec. Where you mount the probe matters more than what the datasheet says. Tape it to the outside wall of a box and you’re measuring the box. Put it in the airflow of a fan and you’re measuring the fan. Thermal mass, self-heating, and air-versus-product-core placement will eat more accuracy than the sensor does.
A cold room is a metal box full of water. Metal and liquid are the two worst attenuators for 2.4 GHz radio. Whatever range figure you were quoted, halve your expectations indoors and plan gateway density from there. This is the single most common reason a pilot that worked in an office fails in a warehouse.
Real world: a cold chain that stopped arguing about proof
Picture a mid-sized third-party logistics operator in Rotterdam — chilled and frozen food, four cold rooms, about thirty reefer trailers, plus a small pharmaceutical line. Their problem wasn’t finding pallets or tracking trailers. It was a rejected shipment of soft fruit, worth roughly €40,000, where the retail buyer claimed a temperature excursion in transit. The 3PL had a handwritten probe log and a driver’s word. They lost the argument and ate the cost.
What they deployed was hardware only, chosen to match four different jobs:
- S23 loggers taped inside outbound pallets on every retail load — cheap enough to treat as consumable, and each one produces a downloadable PDF or Excel record of the whole journey.
- S6et2 probe sensors screwed to the walls of each cold room as fixed reference points, running on their 8-year batteries so nobody has to think about them again.
- KG02 gateways mounted in each room, and a KG01 at the loading docks where the Coded PHY range covers the yard approach.
- LAS6 LoRaWAN sensors on the outdoor container plug points, where there’s no cabling and no Bluetooth coverage worth the name.
Everything pushed over MQTT into the WMS and quality platform the client already owned. The next time a buyer disputed a load, the quality manager pulled the pallet’s log, attached the exported report, and closed it in an afternoon instead of three weeks. The manual probe rounds stopped, because the fixed sensors were already doing that job continuously.
Nothing in that story is exotic. That’s rather the point. The hardware did one thing — record conditions reliably and hand them over — and the client’s own software did the thinking.
The part about who owns the software
Worth stating clearly, because it changes how you should evaluate the whole category: KKM builds hardware. Devices, firmware, OEM and ODM work, white-labelling, certifications. They ship API and SDK documentation so the hardware talks to your system. They don’t sell a monitoring platform and they don’t hold your data.
For a logistics operator with a WMS, a pharma company with a validated quality system, or an integrator building for a client, that’s the arrangement you want. Alert thresholds, escalation rules, dashboards, audit exports — that’s your logic and your IP, and it belongs in your stack, not in a hardware vendor’s subscription. KKM has been doing this since 2008, with 50+ patents and an ISO9001-certified factory, which is exactly the kind of boring you’re shopping for here.
Before you spec anything
Five questions that separate a working deployment from a pilot that dies at month three:
- What interval do you actually need, and what does it cost in battery? Multiply it out across your device count before you commit.
- How much does it store locally, and what happens when the network drops? If it can’t hold data offline, you’ll lose exactly the events that matter.
- Where does the data land? Confirm the gateway pushes to your endpoint over open protocols, and that nothing in the chain quietly needs a vendor subscription.
- What’s the accuracy at your operating temperature? Manufacturers quote their best window. A logger rated ±0.4°C from 0-60°C may only be ±0.6°C below zero — and that’s before mounting effects.
- Does the export survive an audit? Timestamped PDF or Excel with the full record, or a screenshot of a dashboard? Only one of those has ever won an argument.
The short version
Location tracking matured. It’s cheap, it’s accurate, and it’s no longer a reason to choose one vendor over another. What’s left is condition — the temperature a shipment reached, the humidity in a store room, the events a facility can prove happened and when.
That’s a hardware problem, and it’s mostly about being boring in the right ways: long battery life, generous local storage, open protocols, and specs that hold up in a metal box full of water. Pick the device for the job, keep the software under your own roof, and the audit trail takes care of itself.
Sources
- IoT Insider — IoT in 2026: the technologies driving the next wave of IoT growth (1 January 2026)
- The Network Installers — IoT Device Growth Statistics & Trends (2026) (12 January 2026)
- Semtech — LoRaWAN reaches critical mass: 125 million devices and accelerating growth
- KKM — S23 Cold Chain Temperature Logger
- KKM — K4PT BLE Temperature Data Logger
- KKM — S6e Bluetooth Temperature and Humidity Sensor
- KKM — LAS6 LoRaWAN Temperature Humidity Sensor
- KKM — KG01 Outdoor Bluetooth Gateway
- KKM — KG02 Bluetooth Beacon Gateway
- KKM Smart Solutions — Bluetooth Low Energy Beacons
