Energy efficiency in IoT is today one of the most important design parameters — it determines whether a device will run on a single battery for a few weeks or for over a decade. In distributed sensor fleets, where replacing a cell means a technician's visit to the field, every microamp of quiescent current translates directly into maintenance costs and the real return on a deployment. Designing for low energy consumption is therefore not an "end-of-project" optimisation, but an architectural assumption spanning hardware, firmware, and cloud.
In short: energy efficiency in IoT is achieved by combining deep microcontroller sleep modes (a current of 1-10 µA), infrequent transmission over LPWAN networks (LoRaWAN, NB-IoT), and a well-thought-out energy budget — as a result, a sensor on a 2,400 mAh lithium battery can run for 5 to 15 years.
What is energy efficiency in IoT and why is it crucial?
Energy efficiency in IoT is a device's ability to perform its measurement and communication functions with minimal power consumption, most often measured in microamps (µA) and millijoules (mJ) per cycle. In practice this means maximising the time a node spends in sleep mode and minimising the energy consumed during the short windows of activity.
The importance of this parameter grows with the scale of deployment. A fleet of thousands of custom IoT devices scattered across a plant or city makes manual battery replacement costly and logistically difficult, which is why low energy consumption directly lowers the total cost of ownership (TCO).
How do microcontroller sleep modes extend battery life?
Sleep modes are low-power states in which the microcontroller shuts down clocks and peripherals, leaving active only the circuits needed to wake up. It is these, not the processor in active mode, that determine battery life — a typical sensor device spends over 99% of the time asleep.
The differences between chips are enormous. An ESP32 in deep sleep draws about 10 µA, an STM32L in Stop mode drops below 1 µA, while the same chip in active mode with an active radio can draw 100-240 mA. We described the choice of hardware platform in more detail in our comparison of ESP32 versus STM32 for IoT projects.
Key firmware techniques that reduce energy consumption:
- Duty cycling — the device wakes only for the duration of measurement and transmission, e.g. for 200 ms every 15 minutes.
- Tickless idle in the RTOS — putting the core to sleep between events instead of a cyclic timer interrupt; more in the article on FreeRTOS in IoT firmware.
- Interrupt-driven wakeup instead of polling, which keeps the processor in an active state.
- Shutting down unused peripherals (ADC, UART, GPIO pull-ups) during sleep.
Which radio protocol consumes the least energy?
The radio is usually the most energy-hungry element of an IoT device, which is why the choice of protocol has a greater impact on battery life than the microcontroller itself. The rule is simple: the less frequent and shorter the transmission and the lower the draw in listening mode, the longer the battery operation.
For infrequent sending of small telemetry packets, LPWAN networks work best. Our comparison of LoRaWAN, NB-IoT, and LTE-M for asset tracking shows that LoRaWAN with class A allows for years of operation, while NB-IoT with PSM and eDRX modes drastically reduces modem consumption between transmissions.
Approximate draw in transmit mode: LoRaWAN 20-130 mA (short window), NB-IoT 100-300 mA, Bluetooth Low Energy 5-15 mA, classic Wi-Fi 150-300 mA over a longer window. For local communication a common choice is BLE — we described the energy trade-offs in our comparison of BLE versus Wi-Fi in IoT devices.
How do you design the energy budget of an IoT device?
The energy budget is the balance of all energy consumed in one full operating cycle of the device, expressed in millijoules or microamp-hours, referenced to the cell's capacity. It is the fundamental tool that lets you estimate battery life even before building a prototype.
We estimate the projected operating time by dividing the usable battery capacity (in mAh) by the average current draw, determined as a time-weighted average of the individual states. Example: with a sleep current of 5 µA, an active draw of 120 mA for 0.3 s every 15 minutes, and a 2,400 mAh battery, the average draw is about 45 µA, which gives over 6 years of operation.
At the PCB stage, leakage currents and power-supply efficiency also matter — which is why designing a PCB for IoT includes selecting LDOs or converters with low quiescent current and power gating.
How do edge computing and TinyML reduce energy consumption?
Processing data on the device is energetically cheaper than sending raw measurements to the cloud, because radio transmission costs many times more energy than local computation. Filtering and aggregating data "at the network edge" reduces the number and size of packets.
Instead of sending hundreds of samples per second, the device can send one aggregated value or just the anomaly event itself. This approach ties in with edge computing in IoT and with the lightweight inference of machine-learning models described in the article on TinyML on the ESP32-S3, where classification happens without waking the radio link.
Energy harvesting: when should you give up the battery?
Energy harvesting is the acquisition of energy from the environment to power or recharge an IoT device without replacing the cell. It makes sense where a stable energy source is available and the node's energy budget is low enough to be covered by the harvested power.
The most common sources are photovoltaics (from a few dozen µW/cm² indoors to mW/cm² outdoors), thermogenerators (TEG) exploiting a temperature gradient, and piezoelectric transducers recovering energy from machine vibration. Implementation requires a PMIC circuit with an energy store (a supercapacitor or a small battery) and firmware that matches activity to the available power.
Frequently asked questions (FAQ)
How many years can an IoT sensor run on a single battery?
A well-designed IoT sensor on a lithium battery (e.g. ER14505, 2,400 mAh) can run for 5 to as many as 15 years. This is determined by the current in deep sleep mode (on the order of 1-10 µA), the infrequency of transmission, and the choice of an energy-efficient radio protocol such as LoRaWAN or NB-IoT.
Which radio protocol is the most energy-efficient?
For infrequent sending of small packets, the most energy-efficient are LPWAN networks: LoRaWAN and NB-IoT. Bluetooth Low Energy works well for local communication, while classic Wi-Fi consumes the most energy and is rarely suitable for devices powered by a battery for many years.
What is energy harvesting in IoT?
Energy harvesting is the acquisition of energy from the environment — light (photovoltaics), vibration, a temperature gradient, or radio waves — to power an IoT device without replacing the battery. It allows for the building of maintenance-free sensors, but requires a precise energy budget and power management circuits (PMIC).
Summary and key takeaways
Energy efficiency in IoT is the result of consistent decisions at all layers: the choice of a microcontroller with low sleep current, an energy-efficient radio protocol, a solid energy budget, and processing data at the network edge. Neglecting even one layer can shorten battery life from years to months.
At FSS we design IoT devices comprehensively — from schematic and PCB, through energy-efficient firmware, to cloud backend — so that battery life meets the business assumptions of the deployment. If you are planning a fleet of sensors with multi-year autonomy, learn about our approach to building custom IoT devices and consult us on your project's energy budget.
{“@context”: “https://schema.org”, “@type”: “Article”, “headline”: “Energy Efficiency in IoT: How to Extend Battery Life”, “about”: “Energy efficiency in IoT, sleep modes, LPWAN, energy budget, energy harvesting”, “inLanguage”: “en-US”, “author”: {“@type”: “Organization”, “name”: “FSS”, “url”: “https://fss.cc/”}, “publisher”: {“@type”: “Organization”, “name”: “FSS”}, “keywords”: “energy efficiency in IoT, low energy consumption, sleep modes, LPWAN, energy budget, energy harvesting”}
{“@context”: “https://schema.org”, “@type”: “FAQPage”, “mainEntity”: [{“@type”: “Question”, “name”: “How many years can an IoT sensor run on a single battery?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “A well-designed IoT sensor on a lithium battery (e.g. ER14505, 2,400 mAh) can run for 5 to as many as 15 years. This is determined by the current in deep sleep mode (on the order of 1-10 µA), the infrequency of transmission, and the choice of an energy-efficient radio protocol such as LoRaWAN or NB-IoT.”}}, {“@type”: “Question”, “name”: “Which radio protocol is the most energy-efficient?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “For infrequent sending of small packets, the most energy-efficient are LPWAN networks: LoRaWAN and NB-IoT. Bluetooth Low Energy works well for local communication, while classic Wi-Fi consumes the most energy and is rarely suitable for devices powered by a battery for many years.”}}, {“@type”: “Question”, “name”: “What is energy harvesting in IoT?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Energy harvesting is the acquisition of energy from the environment — light (photovoltaics), vibration, a temperature gradient, or radio waves — to power an IoT device without replacing the battery. It allows for the building of maintenance-free sensors, but requires a precise energy budget and power management circuits (PMIC).”}}]}