Skip to main content

Ah, the environment. It used to be just where we happen to live but these days it's a political hot potato that can rend family gatherings more adroitly than Jason Vorhees. Unfortunately, when topics become that polarising, it can be extremely difficult to chart the most pragmatic path towards the best outcome.

In this article, we will consider both the environmental impact of the Internet of Things (IoT) technology and some of the ways it can help mitigate environmental pressures to try and present something approaching a balanced view of the subject.


All human activity has an environmental impact but determining the scope of that impact can be quite challenging. For example, when we consider the environmental impact of AI, we might think of the energy required to build the electronics that make up a server farm. We might even consider how environmentally friendly the mining operations are for extracting the elements used to make electronic components. But what about training an AI system?

MIT Technology Review released a study in 2019 that found that training what you might call a standard AI, that uses a single high-performance graphics card, has the same carbon footprint as a flight across the United States. When training a sophisticated AI like a natural language processing (NLP) application, the carbon footprint is five times greater than the entire life cycle of an American car, including its manufacturing. That’s quite a lot for an all but invisible environmental factor.

When it comes to IoT the most obvious impact factor is its sheer scale. Depending on whose guestimates you are using, there were anywhere between 31 billion and 36 billion IoT devices in the wild at the end of 2021. Projections for 2025 are for there to be around 75 billion devices, rising to 125 billion devices by the end of the decade. That’s a lot of landfill.

There are two issues specific to 'smart' devices that exacerbate the already huge problem of e-waste. Manufacturers are adding semiconductors to products that previously had no need of them, which in turn is shortening the life of these devices as more computing is added: meaning that products that used to have a 15-year lifespan are now being replaced in under five years.

It is also a reality that many small connected devices such as trackers, jewellery and wearables are considered failed once the battery dies. At that point, the consumer expects to bin it and buy another.

Maybe this wouldn’t be so much of a concern, but as Designspark’s own esteemed Bill Marshall has previously pointed out, there is already rather a lot of e-waste. When I say a lot, try an estimated 57.4 million tonnes just in 2021 – that’s more than the weight of the Great Wall of China, the world’s heaviest artificial object.

The Circular Economy

So, if we are going to be installing all this new infrastructure, how can the impact be mitigated? One suggestion gaining a lot of traction with policymakers is the idea of moving to a circular economy. This is, in essence, an economy where material resources are never wasted but continually reused.

Circular economy graphic

In theory, this economic model eliminates, or at least minimises, the extraction of new materials from the earth and eliminates the waste going to landfill or incineration. Amsterdam has declared that it will become a fully circular economy city by 2050, and major corporations like Volkswagen and Unilever have touted their intentions to work towards a circular economy.

While ‘green’ proposals like this are headline-grabbing, the idea itself has already been recycled a few times and was initially proposed in the 1980s by Daniel Knapp, who spread the idea of ‘Total Recycling’ and coined the term 'Zero Waste' in response to the growing trash mountain from 1980's consumerism.

How does a circular economy differ from standard recycling? Good question. The main problem with standard recycling is that materials become contaminated and degraded during the recycling process. This makes it impossible for them to be reused in the same quality applications, so the materials become less useful every time they are recycled. They also tend to be in the wrong place to be reused. For example, the materials in an item made in China, but sold in the UK will be in the wrong location to be used again, even if they could be recycled without degradation.

To achieve something approaching circularity, we would need to design systems so the quality of materials can be maintained over an infinite number of repeated cycles and where recycling and production happen in close proximity. This runs counter to the current globalisation trend, perhaps needing extreme localization of manufacturing and remanufacturing. It also runs counter to the very real physics of dissipation and entropy where every loop around the circle requires new materials and energy to overcome dissipative losses. Good academic discussions of the practical limits to circularity can be found here and here.

The Electronics Industry

Electronic equipment is arguably one of the least circular manufacturing sectors, with so many types of rare and unusual materials bonded together in horrendously complex products that are extremely difficult to separate and recycle. That said, some companies are making the attempt at circularity.

Spire make an adhesive wearable health tracker with a nonreplaceable internal coin cell battery that dies after about 18 months. When this happens, the expired tag can be returned to Spire, who designed each of the components inside the device to be easily taken apart for recycling. It's no mean feat to find glues that allow a wearable to be machine-washed. Making something washable, waterproof and easy to disassemble takes a lot of engineering.

That begs the question of whether, outside the medical sector, industry and consumers are ready to break from cheap, disposable items in favour of more expensive items that have been rigorously designed for maximum recyclability? You also need to factor in the expense of recovering failed units, especially from remote deployments. I’m sure everyone will agree it’s the right thing to do, but people have a habit of saying one thing and doing another, especially when confronted by a financial disincentive.

Self Mitigation?

The largest elephant in the room for the circular economy is that, even with the best will and intent, it is still a long way off. The next decade’s worth of landfill is being manufactured and deployed today. That being the case, can existing IoT do more in the way of good for the environment while deployed than any damage it may cause? There are those who would answer that question with an enthusiastic ‘Yes!’

The white paper by 6GWorld proposes that by 2030, IoT could save approaching 1.8 PWh (a PetaWatt is 1015 watts) of electricity usage and an additional 3.5 PWh of hydrocarbon fuel use for a total energy saving of 5.3 PWh or around a gigaton of reduced CO2 emissions. This is against 653TWh (a TeraWatt is 1012 watts) used to power the IoT units making these savings. This is mostly from smart building management, fleet management and traffic monitoring and control.

Part of the proposal that may become more pressing over time is conserving nearly 230 billion cubic meters of freshwater, especially when it is projected that 52% of the world's population will be living in water-stressed regions by 2050. 35% of the saving is projected to be from improved water grid operations and the larger share from IoT-enabled agricultural crop management.

You get all of these proposed benefits for a measly 657,000 tons of extra eWaste.

Other Uses

Outside of the well-worn tropes for IoT, like building management, traffic management and agriculture, people are finding other uses for IoT that have direct effects on the environment.

One is in the food supply chain. Nearly one-third (1.6 billion tons) of the total food produced globally is wasted. One way to reduce this staggering waste is using cold chains controlled by wireless IoT sensors. They can monitor ambient conditions across the distribution chain to maintain optimum light intensity, air quality, humidity and temperature; ensuring the food remains fresh until it can be consumed.

Another set of uses is in direct environmental preservation:

A cutting-edge captive breeding centre called La Olivilla in Southern Spain has turned around the fortunes of the near-extinct Iberian Lynx: there are now over 300 of these felines. They are being reintroduced into safe habitats and tracked with location collars that georeference each cat like any other IoT asset management system. Connected drones also monitor them visually to see how well they are doing from a distance.

Iberian Lynx

Meanwhile, Rainforest Connection, a San Francisco-based startup, is trying to stop illegal logging and wildlife poaching using innovative IoT systems.

Final Thoughts

Can IoT have a positive impact on the environment? Perhaps.

In the absence of a circular economy, some manufacturers are designing equipment for longer deployment by using ultra-low-power devices combined with energy harvesting technology to lengthen battery lifespans or eliminate batteries altogether. Designing for remote device management and over-the-air (OTA) updates also helps fend off obsolescence and keeps electronic equipment out of landfill for longer.

There are certainly efficiencies to be gained in infrastructure control, especially in industrial processes.

However, the most important factor is us, the consumers. At some point, we have to decide what actually needs to be ‘smart’. This isn’t just the obvious disposable tat. It includes reassessing many of the applications that seem to be a shoo-in for IoT. For example, is ‘smart’ lighting and HVAC actually environmentally better overall than encouraging a responsible corporate culture where the last one out of a room turns off the lights and A/C? Do we really want to encourage the abrogation of personal responsibility with ‘smart’ tech, to the point where most of the human race become Eloi in a generation or two? I don’t profess to know the answers, but questions like these will need to be courageously tackled by policymakers if we’re not to end up living on WALL-E’s trash mountain.

Mark completed his Electronic Engineering degree in 1991 and worked in real-time digital signal processing applications engineering for a number of years, before moving into technical marketing.