IoT sensors for commercial farms: what to install and where.
Sensors are the nervous system of a modern farm. They measure what humans cannot track manually: continuous temperature, moisture, weight, flow, position, and composition data across your entire operation, 24 hours a day. But most farms that have sensors treat them as standalone gadgets. A temperature probe here, a weather station there, a GPS tracker on two of the six trucks. Each device doing its job in isolation. None of them connected to anything else.
A 2025 Stellenbosch University and BFAP national survey found that while 43% of SA field crop producers report full precision agriculture adoption and another 51% partial adoption, the underlying data infrastructure is fractured. Of farmers who collect data, 45% still use paper records and 10% use spreadsheets, according to SA Grain Magazine. Sensors without a system are just expensive thermometers.
This guide covers the major sensor categories for commercial farms in South Africa, what each type actually costs, where to install it, and what decisions it enables when connected to a central operating system.
The sensor categories that matter
Not every farm needs every sensor. The right mix depends on your operation. A dairy farm has different priorities than a poultry house or a grain operation. But the categories below cover the full spectrum. Pick the ones that match your business.
1. Milk meters and conductivity sensors
What they measure: Per-cow milk yield (litres per milking), milk conductivity (electrical resistance), flow rate, and milking duration.
Where they go: Installed inline at each milking point in the parlour. The three main brands available in South Africa are Afimilk (distributed through Waikato SA), DeLaval, and GEA. Lely systems include integrated sensors as part of their robotic milking platforms.
Decisions they enable: Individual cow performance tracking. Early mastitis detection through conductivity anomalies, catching subclinical cases that cause 10 to 25% yield reduction before visible symptoms. Identification of underperforming animals. Lactation curve analysis for breeding and culling decisions. SA processors typically penalise at somatic cell counts above 400,000 cells/mL, with bonuses below 300,000. Continuous monitoring keeps you on the right side of that line.
2. Bulk tank probes
What they measure: Milk temperature, fill level, agitation status, and cooling performance over time.
Where they go: Submerged in the bulk tank, wired to a monitoring unit that transmits data continuously.
Decisions they enable: Real-time quality assurance. Immediate alerts if temperature rises above safe thresholds. This is critical in South Africa: load shedding cost SA agriculture R23 billion in 2022, according to the USDA Foreign Agricultural Service. If bulk tank cooling fails for more than two hours, spoilage risk escalates rapidly. A connected probe triggers alerts before that window closes, giving you time to switch to generator backup. Fill rate tracking against expected herd output detects when production is below expectations. Prevention of costly load rejections at the processor.
3. Soil moisture probes
What they measure: Volumetric water content at different soil depths (typically 10cm, 30cm, and 60cm). Some also measure soil temperature and electrical conductivity (salinity).
Where they go: Buried at multiple depths in representative locations across irrigated lands or pastures. One sensor cluster per zone, positioned to represent the dominant soil type in that zone.
SA brands and pricing: Sentek (Australian, widely used in SA), CropX, and Davis Instruments are common options. Budget R3,000 to R15,000 per probe unit, depending on depth capability and connectivity features. Water monitoring is especially critical in South Africa, where borehole level sensors, dam level monitoring, and flow meters for irrigation can prevent catastrophic water waste in the water-scarce Western and Eastern Cape.
Decisions they enable: Precision irrigation scheduling. Reduced water waste. Root zone monitoring to prevent waterlogging or drought stress. Correlation with yield data to identify optimal moisture ranges for specific crops or pastures.
4. Weather stations
What they measure: Temperature, humidity, wind speed and direction, rainfall, barometric pressure, solar radiation, and evapotranspiration.
Where they go: Mounted on a mast in an open area on the farm, away from buildings and trees. Ideally at standard meteorological height (2m for temperature/humidity, 10m for wind). Larger farms may need multiple stations.
SA brands and pricing: A Davis Vantage Pro2 runs R15,000 to R25,000. Campbell Scientific stations, used in research-grade deployments and by the Agricultural Research Council (ARC), start at R50,000 and go up from there. For most commercial farms, the Davis is more than sufficient.
Decisions they enable: Spray window identification (wind speed, humidity, temperature thresholds). Frost alerts. Irrigation scheduling in conjunction with soil moisture data. Harvest timing. Heat stress management for livestock and poultry. Historical weather data for insurance claims and long-term planning.
5. Feed scales and load cells
What they measure: Weight of feed in storage silos, mixer wagons, and feed bunks. Some systems also track individual component weights during ration mixing.
Where they go: Under feed silos (load cells at the base), on mixer wagons, and at feed bunk delivery points.
Decisions they enable: Accurate feed inventory management. Ration accuracy verification (was the mix actually what was prescribed?). Feed waste on poorly managed farms runs 5 to 15% of total feed cost through inaccurate mixing, spillage, and poor storage. For a 300-cow dairy spending R300,000 to R500,000 per month on feed, 10% waste is R30,000 to R50,000 per month, gone. Connected scales catch that drift in days, not months. Automatic reorder triggers when silo levels drop below threshold. Feed-to-yield ratio calculations when connected to production data.
6. Energy monitors
What they measure: Electricity consumption per circuit or per piece of equipment. Voltage, current, power factor, and demand profiles over time.
Where they go: CT clamps installed on main distribution boards and on individual circuits for high-consumption equipment (cooling systems, pumps, milking equipment, grain dryers, poultry ventilation).
Decisions they enable: Dairy cooling systems typically consume 30 to 60 kWh per day for a 200 to 400 cow operation. On Eskom's agricultural Landrate tariff at R1.80 to R2.50 per kWh, that is manageable. But during load shedding, diesel generation costs R5 to R15 per kWh, doubling or tripling energy spend. Energy monitors identify which equipment to prioritise on generator, where solar plus battery solutions can offset demand, and when equipment is running outside normal parameters (a motor drawing more current than normal often indicates a mechanical problem before it fails).
7. Climate sensors (poultry and controlled environments)
What they measure: Air temperature, humidity, CO2, ammonia (NH3), airflow velocity, and litter moisture in enclosed production environments.
Where they go: Multiple points inside poultry houses, piggeries, or any enclosed livestock facility. Typically at bird level and at exhaust points. Multiple sensors per house to capture temperature gradients. Temperature loggers run R500 to R3,000 per unit.
Decisions they enable: Ventilation and heating system optimisation. Early detection of environmental stress (high ammonia, high CO2, temperature extremes). SA broiler mortality rates run 4 to 7% per cycle, with each bird lost costing R40 to R80 depending on age at death. Precise climate control reduces that mortality rate significantly. Energy savings through targeted ventilation rather than running fans at full speed continuously.
8. GPS and fleet trackers
What they measure: Real-time position, speed, engine hours, fuel consumption, and geofence events for vehicles and mobile equipment.
Where they go: Hardwired into each vehicle or mobile asset: tractors, trucks, bakkies, ATVs, trailers, and irrigation pivots. GPS trackers run R2,000 to R8,000 per unit.
Decisions they enable: Fleet utilisation analysis (which vehicles are idle, which are overused). Fuel theft and waste detection. Maintenance scheduling based on actual engine hours. Route optimisation. Field operation tracking (which tractor sprayed which field, when, and for how long). Labour verification. SA-specific consideration: theft and vandalism of equipment is a real risk, so hardwired trackers with tamper alerts provide both operational and security value.
Connectivity: getting the data off the farm
Sensors are only useful if their data reaches a system that can process it. On a commercial farm, connectivity is often the hardest problem. Only 1.7% of rural South African households have home internet access, according to ITWeb. Not every paddock has cell signal. Not every shed has Wi-Fi. Here are the realistic options for South African farms.
- LoRaWAN (Long Range Wide Area Network). Low-power, long-range wireless protocol. A single gateway on the farmstead can cover 5 to 15 km in open farmland depending on terrain. Ideal for battery-powered sensors in remote locations (soil moisture, water tank levels, borehole monitors, gate sensors). The Things Network has SA gateways already operational. IoT.nxt, a Pretoria-based company acquired by Vodacom, builds enterprise IoT infrastructure locally. Low data throughput, so it suits sensors that report small packets every few minutes.
- LTE/4G cellular. Vodacom and MTN cover most commercial farming areas, but not all remote locations. Weather stations, GPS trackers, and camera systems often use cellular connections. Requires a SIM card and data plan per device or gateway.
- Starlink and satellite. Starlink launched in South Africa in 2024. It costs approximately R4,500 per month plus R12,000 for hardware. Expensive compared to cellular, but it delivers broadband connectivity anywhere in the country. For farms in deep rural areas with no cellular coverage, it can serve as the farm's primary internet connection, with all sensors routing through the local network to the Starlink terminal.
- Mesh networking. Sensors within a facility (poultry house, milking parlour, processing area) can form a local mesh network that connects to a central gateway. LoRa-based mesh networks cover 5 to 15 km per gateway in open farmland. Useful for dense sensor deployments in a confined area.
Most commercial farms end up with a combination: LoRaWAN for remote field sensors, local networking inside buildings, and LTE or Starlink as the backhaul to the cloud. SA-specific challenges to plan for: dust, heat, and humidity damage to exposed hardware, power supply in remote areas (solar plus battery is the standard solution), and theft or vandalism of equipment in the field.
The critical mistake: sensors in isolation
Here is where most farms go wrong. They buy a weather station from one vendor, a GPS tracker from another, soil sensors from a third, and a milking system from a fourth. Each system has its own app, its own login, its own dashboard. None of them share data. None of them know about each other.
A soil moisture sensor that does not know the weather forecast will tell you the soil is dry today. It cannot tell you that 30mm of rain is coming tomorrow, so you should hold off on irrigation. A GPS tracker that does not know the spray schedule will tell you where the tractor is. It cannot tell you whether the spray window has closed because the wind just exceeded threshold.
Sensors in isolation produce data. Sensors in a system produce intelligence. The value is not in any single reading. It is in the connections between readings, across domains, in real time. McKinsey estimates that connected agriculture could add $500 billion or more to global GDP through a 7 to 9% improvement in farm-level productivity. That improvement comes from connection, not collection.
This is why every sensor deployment should start with the question: where does this data go? If the answer is "into its own app that nobody else can access," you are building another data silo. If the answer is "into a central operating system that connects it with financial, operational, and production data," you are building something that compounds.
Getting started
You do not need every sensor on day one. Start with the systems that address your biggest pain points. For dairy, that is usually milk meters and bulk tank monitoring. For poultry, climate sensors. For crops, soil moisture and weather. For every operation, energy monitoring and GPS fleet tracking tend to pay for themselves fastest.
The total sensor investment for a mid-sized commercial farm is meaningful but not prohibitive. A weather station (R15,000 to R25,000), a set of soil probes (R3,000 to R15,000 each), GPS trackers (R2,000 to R8,000 per vehicle), and temperature loggers (R500 to R3,000 each) can be deployed incrementally. The important thing is to install them into a connected system from the start. Not as standalone gadgets. Not with plans to "integrate later." Integration later never happens. Build the infrastructure right from day one, and every sensor you add from that point forward makes the entire system more valuable.
Ready to connect your farm?
Request a quote and we will scope a sensor strategy and operating system built around your operation.
Request a quote →← Back to blog