The green and digital “twin transitions” offer the promise of leveraging digital technologies to reach environmental sustainability goals. Digital technologies and their underlying connectivity can significantly accelerate green transformation across economic sectors, but their environmental footprint must also be considered. This chapter focuses on key sectors where policy makers and technology providers can harness digital technologies to help meet environmental sustainability goals and describes ways to mitigate the negative environmental impact from digital technologies. It concludes by outlining policy priorities and the challenges of aligning the green and digital twin transitions to accelerate action for the good of the planet.
OECD Digital Economy Outlook 2024 (Volume 2)
Strengthening Connectivity, Innovation and Trust
Abstract
The world is facing a climate emergency: Will digital technologies help or hinder the green transition?
As the time window for successful climate action narrows, it is increasingly clear that the digital and green “twin transitions” must be harnessed to rapidly decarbonise economies and to achieve the Sustainable Development Goals (SDGs).
However, digital technologies themselves have an environmental footprint along their life cycle, with information and communication technologies (ICTs) making up between 1.5% and 4% of global greenhouse gas (GHG) emissions in 2020 (Bieser et al., 2023[1]).
About 90% of electricity used by data centres is estimated to be lost as waste heat, representing a largely untapped opportunity to apply circular economy models through applications like district heating (Luo et al., 2019[2]).
Digital technologies offer a viable pathway to decarbonisation across sectors
Digital technologies are major building blocks to help achieve the deep cuts in emissions needed for a net-zero emissions world. Technologies like the Internet of Things (IoT) and digital twins enabled by artificial intelligence (AI) can improve energy efficiency, reduce costs, and accelerate innovation across energy grids and product supply chains.
Communication infrastructures and services are fundamental to a sustainable and resilient digital transformation. For example, they are needed to deploy technologies like smart electrical grids to IoT-based precision agriculture, which support decarbonisation of many sectors of the economy. At the same time, they have their own environmental footprint to be minimised.
Sectors including global transportation systems stand to benefit from digital technologies that support reducing environmental impacts through fuel efficiency gains, predictive maintenance and shared mobility, as well as by enabling low-carbon transport systems and multimodal mobility services.
Towards policies for a green and resilient digital future
Digitalisation and environmental sustainability are increasingly being considered together on the agendas of policy makers, including in national strategies and recovery plans, as well as in dedicated strategies for digital technologies and the environment.
The road to a green digital world includes difficult policy questions. For example, as digital technologies increase the demand for computing power and data centres, questions arise around whether local energy grids are ready to support increased green digital transformation.
Countries are amid two fundamental transformations in the 21st century: the transition towards an environmentally sustainable and carbon-neutral economy, and the proliferation of digital technologies in almost all areas of public and private life. Ongoing digital transformation of economies and societies holds many promises to spur innovation; improve productivity and services; connect billions of people worldwide, including in emerging economies; and to generate efficiencies (OECD, 2019[3]), such as resources and productivity. Increasingly, these “green” and “digital” transitions are inter-linked by policy makers and stakeholders. They recognise the importance of ensuring the “twin transitions” are aligned and leveraged for a sustainable future. Adopted over a decade ago, the OECD Recommendation on Information and Communication Technologies and the Environment pioneered work in examining the relationship between digital technologies and the green transition (OECD, 2010[4]). The Recommendation is undergoing a review of relevance to reflect rapid technological changes and the exacerbating climate crisis.1
Avoiding the most severe impacts of climate change and environmental degradation in the next decades is a global challenge that requires urgent action. At the heart of the twin transitions are digital technologies, such as AI and IoT, along with the underlying connectivity enabling them. Digital technologies can significantly accelerate the sustainable transformation of global energy grids, transportation networks and communication infrastructure. At the same time, digital devices and infrastructure have their own environmental footprint – from the extraction of raw materials to greenhouse gas (GHG) emissions derived from energy consumption.
This chapter examines how countries are aligning the green and digital twin transitions.2 It explores how embracing digital technologies in key sectors supports environmental sustainability goals, and how to mitigate the environmental footprint of digital technologies. It concludes with policy considerations on the road towards achieving a digital and environmentally sustainable future.
As the window for successful climate action narrows, digital technologies offer solutions to reach net-zero targets
Digital transformation offers a multitude of social and economic opportunities for people as citizens, consumers and workers (OECD, 2022[5]). Digital technologies play an increasingly important role in modern human societies – from smartphones and wearable technology to fully automated manufacturing sites and logistical networks. Globalisation and digital transformation have been key forces of productivity gains and innovation in the past decades, spurred in large part by the advent of the Internet and decreasing hardware costs. With smart devices widely embedded across economies, digital technologies define today’s markets, and will likely continue to be foundational to the markets of tomorrow.
However, amid an increasingly digital world, countries globally are also feeling the impacts of the mounting climate crisis. The year 2023 was the warmest year ever recorded, with Antarctic Sea ice coverage hitting a record low (NOAA, 2024[6]). OECD work found that during the 2017-21 period, population exposure to heat stress has been high in many countries across the world. Latin America, the Mediterranean Basin, Australia and the United States have been particularly affected. Costa Rica and Israel, for example, experienced more than 140 days of heat stress per year during the same period (Maes et al., 2022[7]).3 The World Meteorological Organisation (WMO) (2023[8]) estimates that between 2023 and 2027 there is a “66% likelihood that the annual average near-surface global temperature will be more than 1.5°C above pre-industrial levels for at least one year”.
These trends sound the alarm for a world that will likely exceed the 1.5°C temperature target of the Paris Agreement. A growing number of extreme weather events, together with ever more grave scientific reports from the expert community, conclude that the world is in a state of climate emergency. This calls for mobilisation of all stakeholders towards swift and concrete actions to ensure the planet’s environmental sustainability.
As the time window for successful climate action narrows, it is increasingly clear to policy makers that the transformative forces of digital transformation must be leveraged for the rapid decarbonisation of the global economy and to achieve the Sustainable Development Goals. OECD countries have called for leveraging digital technologies in the fight against climate change and aligning digital and green policies (OECD, 2022[9]). The Coalition for Digital Environmental Sustainability (CODES), led by the United Nations Environment Programme, envisions a sustainable planet in the digital age. This would foster a “global movement for digital environmental sustainability” to align, mitigate and accelerate the twin transitions (CODES, 2022[10]). At the European level, the European Union (EU) Green Deal follows an integrative approach for the green and digital twin transitions to reinforce each other and accelerate the achievement of ambitious climate goals by 2050 (Muench et al., 2022[11]).
Increased ICT diffusion creates both positive and negative environmental effects
ICTs form the backbone of digital transformation, providing connectivity, storage, memory, data processing and more. They enable applications like Industry 4.0 and a variety of smart consumer products – from automated vehicles to smart meters (OECD, 2018[12]). ICTs have become pervasive over the past decade, underscoring the importance of aligning the digital and green twin transitions. Between 2011 and 2021, the ICT sector4 grew three times faster than the economy as a whole in OECD countries (OECD, 2024[13]).
Between 2010 and 2023, Internet users globally increased by 160% from 2 billion to 5.3 billion (ITU, 2023[14]). In the OECD area, fixed broadband penetration increased from 24 to 35.8 subscriptions per 100 inhabitants between 2010 and 2023. Meanwhile, mobile broadband penetration tripled during the same period (from 43.3 to 134 per 100 inhabitants) (OECD, 2024[15]). Internet adoption and the increase in broadband availability (measured by subscriptions) are some of the indicators that can be used as proxy measures for mass digitalisation.
While such trends point to increased ICT diffusion throughout economies, they also raise questions. Have environmental impacts of digital adoption also increased? Will positive enabling and systemic effects resulting from digitalisation offset any increases? (OECD, forthcoming[16]). Some examples of the enabling effects of ICTs include the use of IoT and AI to reduce household energy consumption; smart (precision) agriculture to improve crop yields and increase resource efficiency; smart energy grids; and the reduction in congestion through remote working, connected vehicles and smart cities.
Today, more than ever, advances in digital technologies enable development of tools that can assist countries in their journeys towards environmental sustainability (OECD, 2022[5]). From more efficient hardware to smart products that help consumers make greener choices, digital technologies offer one answer to how quickly global economies can achieve net-zero targets. A holistic policy approach to leverage the benefits of the twin transitions requires accurate measurements to understand both the positive and negative environmental impacts of digital technologies.
On the one hand, digital technologies can aid countries to achieve climate goals by creating significant efficiency gains through enabling effects across economic sectors. For example, they can help manage energy systems and achieve emissions cuts needed to meet climate targets. Technologies like AI and IoT can provide the necessary speed and scale to reduce emissions and accelerate innovations in key areas. These include planetary digital twins,5 circular economic models and sustainable consumption habits (CODES, 2022[10]).
On the other hand, digital technologies and their supply chains also have negative environmental impacts along their life cycle, such as raw material extraction, energy and water use (IEA, 2023[17]). Although the ICT sector produces lower GHG emissions than other sectors, their environmental footprint must be thoughtfully managed. Gauging future opportunities and risks around the twin transitions first requires an in-depth understanding of several technologies today, such as IoT, AI and their underlying connectivity infrastructure, which carry great potential to accelerate the green transition.
Broadband connectivity, the Internet of Things and artificial intelligence can provide the speed and scale to rapidly green economies
Digital technologies, or ICTs,6 can be referred to as “both different types of communication networks and the technologies used in them” (OECD, 2007[18]).7 The digital technology ecosystem itself combines cloud computing, big data analytics, AI, blockchain and computing power, among others (OECD, 2019[3]). According to the OECD Recommendation on Information and Communication Technologies and the Environment (OECD, 2010[4]), ICTs can have: i) direct effects on the environment (e.g. through their life cycle of production, transportation, operation, and end-of-life); ii) enabling effects in other sectors (e.g. Industry 4.0, smart energy grids, smart agriculture, electric vehicles, etc.); and iii) systemic effects such as behavioural change (e.g. nudging consumers to make greener choices, rebound effects) (Figure 3.1).
Notes: Examples given are for illustrative purposes and are not exhaustive.
Several key digital technologies are expected to be crucial for a green future. Technologies like IoT, AI and their underlying connectivity, are expected to advance in their capabilities and sectoral applications. They will likely have direct, enabling and systemic effects on environmental sustainability.
Broadband connectivity. Communication networks are a key enabler of digital transformation. High-quality, resilient and affordable broadband connectivity is a prerequisite for several other digital technologies such as AI and the IoT. The communication sector is characterised by fast-paced developments such as new network infrastructure and architecture, the convergence of fixed and mobile networks, and a continuous integration of connectivity into other economic sectors. Communication networks can be a key enabler of environmental sustainability in many fields, while the environmental impact of the infrastructure itself will need to be minimised (OECD, 2022[20]).
Internet of Things. The OECD defines IoT as “all devices and objects whose state can be altered via the Internet, with or without the active involvement of individuals” (OECD, 2018[12]). Connected objects may require the involvement of devices considered part of the “traditional Internet”. However, this definition excludes laptops, tablets and smartphones already accounted for in OECD broadband metrics. IoT falls into two categories. In Wide Area IoT, devices are connected through cellular technology and through Low Power Wide Area Networks. In Short Area IoT, devices use unlicensed spectrum with a typical range up to 100 metres. IoT has several environmental applications such as sensors for agricultural sustainability, smart energy grids and smart home applications (OECD, 2018[12]).
Artificial Intelligence. AI8 underpins some of the most promising technological solutions to today’s global challenges, including addressing climate change. AI applications are enabling efficiency gains across sectors and for both computer hardware and software. However, there are concerns around the energy intensity required to train and use some AI systems, especially the most advanced (OECD, 2022[20]). Today’s advanced AI systems have scaled significantly in size, number of parameters and breadth of training datasets used (OECD, 2024[13]), making them energy intensive to train. Research also notes that compute demands for AI systems, such as processing power, has grown faster than hardware performance. This is especially the case for deep-learning applications such as machine translation, object detection and image classification (Thompson et al., 2020[24]). The rise of generative AI systems at the end of 2022 has also emphasised AI inference (e.g. using AI systems once they have been trained), raising questions around their environmental impact as AI diffusion grows.
What is the environmental footprint of digital technologies?
Measuring the direct effects of digital technologies involves assessing environmental impacts along their life cycle. These impacts are typically measured using a multicriteria life-cycle assessment, often in compliance with internationally recognised standards, such as the International Organization for Standardization (ISO) 14040:2006 and ISO 14044:2006 (Benqassem et al., 2021[25]). Life-cycle assessments usually differentiate between several distinct stages, such as production, transport, operations and end-of-life. This approach is advocated by the OECD Recommendation on Information and Communication Technologies and the Environment (OECD, 2010[4]; OECD, forthcoming[16]).
Although life-cycle categories can vary, the general methodology is widely used by researchers (Hertin and Berkhout, 2001[26]). Examples include “Methodologies for the assessment of the environmental impact of the information and communication technology sector” of 2018 (ITU-T, 2018[27]) and the OECD Recommendation on Information and Communication Technologies and the Environment (OECD, 2010[4]). Environmental impacts can be quantified using impact categories articulated in ISO environmental management life-cycle assessment standards. These include global warming (e.g. GHG concentration causing polar warming and ice melt), toxicity (e.g. causing smog and acid rain) and biodiversity loss (ISO, 2006[28]; ISO, 2006[29]; Mickoleit, 2010[30]).
In a handful of cases, operating digital technologies can produce positive environmental impacts (e.g. re-using excess heat produced by data centres). However, in most cases, their direct impacts have negative environmental impacts along their life cycle (e.g. natural resources consumed while operating digital technologies).
Positive enabling and systemic effects of digital technologies (e.g. indirect impacts such as energy-saving sectoral applications or nudging consumers to make greener choices) can sometimes outweigh the negative direct impacts on the environment from manufacturing, transporting, operating and disposing of ICTs along their life cycle. This makes the measurement of “net” environmental impacts complex. Nevertheless, several measurement frameworks aim to establish common indicators and facilitate a comprehensive environmental impact assessment: the Greenhouse Gas Protocol (2004[31]); first, second and third order effects (Hertin and Berkhout, 2001[26]; EITO, 2002[32]); ISO Environmental management – Life-cycle assessment (ISO, 2006[28]; 2006[29]); ITU-T Rec. L.1450 (09/2018) (ITU-T, 2018[27]), among others.
The operation and production of digital technologies have an environmental footprint
While research typically focuses on measuring GHG emissions from digital technologies, environmental impacts from their production could become a significant sustainability challenge, raising both environmental and human rights concerns. Such upstream supply chain measures are typically called “embodied emissions” or “Scope 3” emissions. They remain mostly voluntary in industry reporting frameworks like the Greenhouse Gas Protocol (WBCSD and WRI, 2004[31]) and the Science Based Targets Initiative, which publishes detailed Scope 3 guidance for the ICT sector (Science Based Targets, 2020[33]). Upstream supply chain impacts for digital technologies include the mining of critical minerals that can be environmentally damaging to extract and that carry supply chain and human rights risks.
Global population growth and rising wealth is projected to nearly double consumption of raw materials by 2060 (Livingstone et al., 2022[34]). Growing demand for digital technologies will play a role in driving this increase. Hardware, such as semiconductors, terminal devices and electric vehicle batteries, require mining of critical minerals at larger quantities than ever before. This may point to future resource constraints as demand grows, yet quantities of natural resources remain largely fixed (OECD, 2019[35]).
The production of ICT products requires a variety of materials, such as iron, steel, plastics, glass and various metals. For most of these materials, ICT products account for less than 1% of global annual usage (Malmodin, Bergmark and Matinfar, 2018[36]; Gupta et al., 2020[37]). However, the sector is a primary user of some critical minerals, for which the demand is expected to sharply increase in the future, as batteries become even more central to economic activities.
Demand is growing for key minerals to feed ICT products, raising concerns about their depletion
Demand for ICT products is contributing to a depletion of key minerals, raising concerns about their sustainability. The manufacturing of electrical and electronic equipment relies on the supply of inputs such as cobalt, gallium, indium, palladium and rare earth elements (REE), as well as tantalum, tin, gold and silver (Chancerel et al., 2015[38]). Compared to global use across sectors, the ICT sector was estimated to use over 80% of the world’s indium, gallium and germanium in 2018. This contributed significantly to the material resource depletion potential of these critical minerals, raising questions around sustainability as demand for ICT products further increases (Malmodin, Bergmark and Matinfar, 2018[36]).
The focus of critical minerals has often been on high-volume categories like battery metals and copper. However, smaller critical minerals like gallium and germanium are characterised by high supply concentration. Consequently, these niche minerals could significantly disrupt supply chains for digital technologies despite their small volumes (IEA, 2023[39]).
Forecasting demand for niche minerals remains challenging due in part to lack of transparency in supply chains. The deployment of clean energy technologies such as photovoltaics – a key technology in solar panels – and batteries is “propelling unprecedented growth in the critical mineral market” (IEA, 2023[39]). In all scenarios, the IEA forecasts demand for critical minerals doubling or more than tripling by 2030 alone (IEA, 2023[39]).
However, projecting the impact of growing demand for digital technologies on the availability of key elements such as REEs, gallium or magnesium remains challenging. Many minerals are required in an extremely pure form for ICT uses, with producers of such pure elements often geographically concentrated. For example, the People’s Republic of China produces 90% of global gallium and germanium output, as well as 70% of most REEs. The small volumes of production and opaque supply chains make it difficult to find production data to better forecast and plan supply and demand needs (UNCTAD, 2020[40]).
Market pressure and competition for critical minerals are expected to intensify as these minerals are essential inputs for green and digital strategies. As a result, jurisdictions like the European Union aim to diversify supply chains. They seek to increase secondary supply through circular economic models and substitute scarce materials where possible (European Commission, 2023[41]).
More investment in recycling, recovery and innovative technologies may help offset The impact of increased demand for rare earth elements
The market outlook for critical minerals points towards significant benefits of policies and programmes to encourage recycling and recovery of REEs and other key minerals used in digital technologies, as well as resource efficiency and circularity (European Commission, 2023[41]). This is especially important as most electronic waste is not collected and only about 1% of REEs are recycled. The recycling of elements such as neodymium – magnets used in hard disc drives, mobile phones, and video and audio systems – often takes less than half of the energy needed compared to their extraction (Geng, Sarkis and Bleischwitz, 2023[42]). Research points to high savings potential related to the recovery of REEs from e-waste products through a circular economy approach. However, data protection and safety regulations can sometimes limit such practices. This can require the destruction and disposal of hardware in ways that do not allow recycling (Willenbacher and Wohlgemuth, 2022[43]).
Innovation has reduced the need for infrastructure to produce digital technologies, but demand for ICTs continues to exert pressure on critical minerals. Digital technologies have experienced a profound “dematerialisation” in recent years. Thanks to hardware and software innovation, these technologies require less physical infrastructure for more efficiency (Wäger, Hischier and Widmer, 2015[44]). However, demand growth for ICTs continues to place pressure on securing a steady supply of hardware inputs, including critical minerals.
When digital technologies are used, their operation produce GHG emissions through energy consumption, depending on the local energy mix of renewable energy and fossil fuels, and often requires large amounts of water, for example, for cooling data centres and other ICT equipment (Li et al., 2023[45]).
While ICT use has grown across sectors, its proportion of global emissions has remained flat largely due to hardware efficiency gains and the dynamics of energy markets (IEA, 2021[46]; 2023[17]). Estimates vary due to different scopes and definitions of the ICT sector. However, the global ICT industry (including terminal device hardware such as televisions) was estimated to make up 1.5% to 4% of global GHG emissions in 20209 (Bieser et al., 2023[1]). While this may point to hardware efficiency gains, dynamics in energy markets may also help explain this trend. Some research indicates that around 70% of all ICT GHG emissions can be attributed to electricity, as opposed to embodied emissions from hardware production early in the life cycle (Freitag et al., 2021[47]).
Such findings suggest that investment in renewable energy production and increasing uptake by the ICT sector could facilitate its rapid decarbonisation of the energy sector. This, in turn, would shift the focus on environmental sustainability concerns further up the ICT supply chain (e.g. to Scope 3 emissions) (Bieser et al., 2023[1]).
Data centres have more workload, but their energy use has remained stable
Despite major increases in the workloads of data centres, their use of energy has remained remarkably consistent. Data centres are a key underlying technology enabling digitalisation. According to the IEA, data centres used about 194 Terawatt hours (TWh) in 2014 or 1% of global energy demand (Figure 3.2). By 2020, they still accounted for this same percentage of global energy demand (IEA, 2021[46]).
Global trends in Internet traffic, data centre workloads and data centre energy use
Notes: Updated November 2021. IEA provides a range for data centre electricity consumption: 240-340 TWh for 2022.
Source: IEA (2021[46]), “Global trends in Internet traffic, data centre workloads and data centre energy use, 2010-2020” based on data from Cisco Global Cloud Index and TeleGeography.
This consistency is striking for two reasons. First, the workload in data centres and Internet traffic have significantly increased. Second, data centres consume significant amounts of electricity, higher even than the electricity consumed by some OECD countries (Figure 3.3). Data centre efficiency improvements and the shift to large hyperscale data centres could help explain why data centre energy use has remained constant.
It is difficult to predict the growth in electricity demand for data centres on a global level. The IEA estimates that data centres consumed between 240-340 TWh globally in 2022, a notable increase from prior years. It also expects global electricity demand for data centres to increase significantly in the next decade – in some scenarios, up to twice current demand. However, such projections remain highly uncertain. They depend on the “pace of deployment, range of efficiency improvements” and other technological trends (IEA, 2024[48]).
At the national level, some governments track energy consumption and projected demand for data centres within their borders. In Denmark, annual data centre energy demand is expected to grow from near-zero in 2020 to 5 TWh by 2025 and 7.5 TWh by 2030 (Danish Energy Agency, 2020[49]). In Ireland, a European hub for data centre operators, data centre energy consumption increased by 144% between 2015 and 2020, accounting for 11% of metered electricity consumed in the country in 2021 (Central Statistics Office, 2023[50]). Median-demand scenarios estimate that this figure could rise as much as 23% by 2030 (EirGrid, 2021[51]).
As data centre workloads and Internet traffic grow, energy grids and water supplies may have limited capacity to meet the growing demand for operating digital technologies. Some countries, including the United Kingdom and Denmark, are considering or imposing moratoria on new data centre construction due to strained national power supplies and rising energy costs (Swindhoe, 2022[52]; Fitri, 2023[53]). Climate change, causing more frequent heatwaves and droughts, can also add stress to power grids and data centres themselves, leading to power outages and concerns around water management (Google Cloud, 2022[54]).
Technologies like AI could further increase the demand for ICT infrastructure, as certain models become larger and more energy intensive to train. Beginning in about 2010, the prominence of an AI method called deep learning dramatically increased the size of machine-learning systems and their compute demands. Recent advances in generative AI systems, such as chatbots, are further raising questions about the energy intensity of AI training and the sustainability of underlying ICT infrastructure.
Electricity consumption in TWh in OECD countries compared to global data centre consumption (left axis) and data centre electricity consumption in TWh per 100 000 inhabitants in OECD countries (right axis), 2022
Notes: IEA provides a range for estimated global data centre electricity consumption: 240-340 TWh for 2022 (light pink stacked bar). This excludes energy used for cryptocurrency mining. No electricity consumption data were found for Iceland and Israel.
Source: Authors’ elaboration based on (IEA, 2023[17]; Enerdata, 2023[55]; Ember, 2023[56]; OECD, 2023[57]).
The IEA expects AI to significantly increase the electricity demand of dedicated AI data centres, which could consume “at least ten times its demand in 2023” by 2026 (IEA, 2024[48]). Satisfying the demand for larger models was partially enabled by transitioning from general-purpose processors, such as Central Processing Units. More specialised processors support more efficient compute execution for certain operations (i.e. requiring less energy, less water for cooling and more computations per unit time).
Today, machine-learning systems are predominantly trained on specialised processors optimised for certain types of operations, such as Graphics Processing Units. In recent years, both governments and private sector companies have shown increasing interest in securing the supply of future generations of chips. These would be designed specifically for AI and promise more energy efficiency (OECD, 2023[58]).
Transporting ICT infrastructure relies on a global network of distribution, freight transportation, handling and storage, with associated environmental impacts
The manufacturing and assembly of ICT infrastructure has a significant environmental impact. The process of enabling digital technologies is embedded into highly globalised and complex supply chains. This global network of distribution comprises freight transportation from manufacturing sites to points of assembly and use, and includes handling and storage.
Such a process generates environmental impacts such as air pollution, oil spills, toxic-waste discharges and acoustic pollution, among others (Crawford, 2021[59]; OECD, 2022[20]). In 2022, energy-related GHG emissions resulting from the transport sector accounted for around a quarter of global emissions (IEA, 2023[60]).
While manufacturing and operations are responsible for most life-cycle emissions, it is unclear whether transport-related emissions will increase or decrease in the future. The life-cycle analysis of various types of hardware, such as computers and data centre infrastructure, often attributes a small percentage of GHG emissions from transport. In many cases, it is measured at below 5% of overall life-cycle emissions.
The manufacturing and operations stages often make up the bulk of life-cycle emissions (Gupta et al., 2020[37]; Kaack et al., 2022[61]). On the one hand, as more business and consumer activities take place on line and teleworking trends persist, the share of environmental impacts from the transportation of digital technologies may decrease. On the other, increased digitalisation could also drive-up market demand for the latest digital products and upgrades. This could, in turn, increase transportation requirements for digital technologies.
Water consumption is an often overlooked environmental impact of digital technologies
The impact of water use to support digital technologies is not well understood due to lack of data. During their operation, digital technologies consume water primarily for cooling systems in data centres. Producing digital technology hardware, such as semiconductor fabrication, can also use large amounts of water. Compared to energy use and GHG emissions, the water resource impacts of digital technologies are poorly understood (OECD, 2022[20]). For example, some researchers estimated that only 33-50% of data centre operators compiled and reported water-use metrics in 2021 (Mytton, 2021[62]; Uptime Institute, 2021[63]).
Taking stock of the water consumed by operating digital technologies is critical for resource management and future sustainability. Climate change and the associated growing incidence of extreme droughts stress global water ecosystems. Indeed, more than 2 billion people worldwide already experience water shortages (United Nations, 2022[64]). While availability of renewable freshwater resources varies considerably, water scarcity is a growing problem in many parts of the world. This results in groundwater depletion and more stress on water availability and quality (United Nations, 2022[64]).
With respect to the ICT sector, water usage is significantly harder to estimate than energy consumption and GHG emissions due to lack of reporting, standards and awareness in the industry. One of the few available evaluations puts the water consumption of United States data centres at less than 1% of total water consumption (Mytton, 2021[62]). However, digital infrastructure such as data centres often compete locally with large water users such as hospitals and the agricultural production. For example, the US data centre industry is a top ten water-consuming industry and often clusters in geographic areas that rely on scarce water supplies (Siddik, Shehabi and Marston, 2021[65]).
Severe weather events like droughts are also creating risks to the stability of upstream digital technology infrastructure production. For example, droughts in Chinese Taipei reportedly affected semiconductor production. This was due to the short supply of highly pure water needed to clean factories and chip components during manufacturing (The New York Times, 2021[66]).
Some research indicates that advanced digital technology applications like generative AI have considerable water footprints. One estimate for training OpenAI’s GPT-3 AI model puts its water usage at between 700 000 and 2.1 million litres of freshwater. This is the equivalent of almost one full Olympic-sized swimming pool depending on the data centre location (Li et al., 2023[45]). According to one estimate, water cooling in data centres uses between 0.2 and 0.8 litres of water per kilowatt-hour used. This amounted to about 120 000 Olympic swimming pools per year for the global data centre industry (Andrews et al., 2021[67]).
Semiconductor and end-user device manufacturing also requires large amounts of water. By some estimates, a typical large chip fab uses up to 37 million litres of water a day (Johnson, 2022[68]). As digital technologies diffuse across sectors and uses, and demands for underlying infrastructure grows, careful water resource management will be needed to minimise the sector’s use of potable water and ease local water stress.
Measurement challenges due to complexity of environmental impacts and rapid technological change
The expected growth in demand for digital services and products requires policy makers to understand and measure their impacts on national and local energy systems, power grids and GHG emission targets. Rapid improvements in energy efficiency and the shift to renewable sources of power have largely limited growth in overall ICT energy demand and GHG emissions. However, the IEA asserts that “strong government and industry efforts on energy efficiency, renewables procurement and research and development (R&D) will be essential to curb energy demand and emissions growth over the next decade” (IEA, 2022[69]).
Methodological challenges for estimating future GHG emissions include a lack of consistency in system boundaries and taxonomies, challenges in data availability and quality, and complexities with measuring enabling and systemic environmental impacts (Bremer et al., 2023[70]). Such challenges are reflected in the divergence of both past estimates and future predictions of the life-cycle GHG emissions of digital technologies. Published estimates for the ICT sector for 2015 “differ by a factor of 2”, while “projections for 2025 diverge up to 25 times” (Bremer et al., 2023[70]).
While operating digital technologies consumes natural resources, their operation (direct effects) can sometimes contribute positively to environmental sustainability. Such “positive” cases have become more technologically viable in recent years, but scaling and commercialisation challenges remain. For example, data centres produce large amounts of excess heat, typically considered “low-grade energy”. This energy usually cannot be directly repurposed for other activities as the temperatures are too low, typically below 35°C, often causing excess heat to be directed into cooling towers.
Several methods have been proposed to recover excess heat from data centre operations. These include combining the operation of a data centre and an onsite greenhouse or transferring it to local energy networks (ReUseHeat, 2017[71]; Karnama, Haghighi and Vinuesa, 2019[72]). This is known as “district heating” (Box 3.1). While such applications seem promising, they will typically require longer-term planning and co-ordination with urban development. They must overcome challenges around transporting such heat to the end-user effectively and related business models.
Re-using waste heat from data centres can enable energy savings and support a circular economy
About 90% of the electricity used by data centres is converted into low-grade waste heat, which is typically lost and released into the atmosphere or discharged into local waterways (Luo et al., 2019[2]). This offers an opportunity to tap into a scarcely used source of energy and support circular economy and local decarbonisation efforts by providing heat to nearby commercial and residential buildings (IEA, 2023[17]). On a small scale, some data centre providers have already started to distribute waste heat to municipal district heating networks.
In Finland, one data centre operator – Elisa – provides heat to 1 000 homes through Helsinki Energy, while QTS Data Centres in the Netherlands heats more than 5 000 homes (Fisher, 2023[73]). District heating works well for urban areas, but there are often limitations to its use in rural areas due to a lack of a critical mass of customers. Nevertheless, even such challenges can be overcome. For example, heated wastewater from the data centre has the optimal temperature for lobsters to grow. With this in mind, a data centre operator in Norway has pioneered heat re-use for land-based lobster and trout farming (Green Mountain, 2021[74]). Countries like Germany have also proposed legislation to make it mandatory for data centre operators to provide waste heat to local suppliers (McGovan, 2023[75]). The IEA recommends policy makers work with data centre operators, utility companies and district heating suppliers to overcome barriers to scale-up and adoption, such as contractual and regulatory challenges, or achieving the required temperature for heating (IEA, 2023[17]).
In addition, while increased hyperscale operations have correlated with increased energy demand overall, such centralisation has also enabled large efficiency gains in energy use. New forms of cooling data centres are emerging but require scaling through substantial upgrades to infrastructure. Data centres based on liquid cooling could also recover the excess heat on-site to heat nearby buildings, in what some call an Organic Data Centre approach (Karnama, Haghighi and Vinuesa, 2019[72]). In addition to energy used directly by data centres, energy is also used to supply treated water to the data centres and treat the wastewater released from them (Siddik, Shehabi and Marston, 2021[65]). However, this wastewater often must be treated after use, which also consumes electricity and produces emissions, rendering the net sustainability impacts unclear (OECD, 2022[20]).
Recycling and end-of-life for digital technologies will need to be further improved to support climate and economic resilience
The ICT sector generates large amounts of electronic waste, which is expected to grow as demand for digital products rise. Waste from electrical and electronic equipment (WEEE or e-waste) refers to discarded electrical or electronic items without the intent of re-use (OECD, 2019[76]). The Global E-waste Statistical Partnership, led by the United Nations and the International Telecommunication Union, estimates that global e-waste amounted to 57.4 million metric tonnes in 2021, up almost 30% from 2014 (Forti et al., 2020[77]). This is further projected to grow to 82 million metric tonnes by 2030 (Baldé et al., 2024[78]). This makes e-waste “one of the world’s fastest growing waste streams”: its growth rate is estimated to be over three times higher than other prominent waste streams (Kumar, Holuszko and Espinosa, 2017[79]).
The Global E-waste Statistical Partnership estimates that only 22.3% of global e-waste was formally collected and recycled in 2022 (Baldé et al., 2024[78]). This leads to environmental concerns as large amounts of e-waste are incinerated or dumped in landfills, leading to environmental and social impacts such as air pollution, acidic and radioactive waste, and groundwater pollution (Crawford, 2021[59]). E-waste can contain more than 100 metals and materials such as lead and other toxic materials such as mercury, lithium and nickel. Consequently, it makes up around 70% of global surface-level toxic pollution (Vishwakarma et al., 2022[80]).
Digital technologies are not directly responsible for the largest amounts of global e-waste. The definition of e-waste is broad and generally comprises six categories: temperature exchange equipment (such as refrigerators and heat pumps), screens and monitors, lamps, large equipment (such as washing machines), small equipment (such as vacuum cleaners and toasters), and small IT and telecommunication equipment (OECD, 2019[76]). In 2019, the largest categories of global e-waste, such as small equipment and temperature exchange equipment, could not be directly attributed to digital technologies. Small IT and telecommunication equipment grew only moderately compared to 2015. Meanwhile, screens even declined by 1%. This was because light-emitting flat panel displays or screens (including light-emitting diodes [LEDs]) led to a decrease in total weight, even as the total number of screens increased (Forti et al., 2020[77]).
There are also significant regional differences in both generation and recycling of e-waste. In 2022, Europe generated 17.6 kg per capita, and recycled 7.53 kg per capita (42.8%). Meanwhile, the Americas10 generated 14.1 kg per capita but recycled only 4.2 kg per capita (30%) (Baldé et al., 2024[78]).
Increasing re-use and recycling rates for ICT infrastructure can help meet the projected increase in demand for critical minerals. The economic value of raw materials contained in e-waste was estimated at USD 57 billion in 2019 (Forti et al., 2020[77]) and rose to USD 91 billion in 2022 (Baldé et al., 2024[78]). A lack of repairs, software support and planned obsolescence strategies for older digital technology products, coupled with low collection and recycling rates, led to high rates of e-waste worldwide. (Forti et al., 2020[77]). The export of significant amounts e-waste for informal disposal in emerging economies also risks creating a ‘leakage effect’ that may have serious consequences.
In terms of recycling, there remains a significant gap across OECD countries between e-waste generated and e-waste that is formally collected and recycled (Figure 3.4) (OECD, 2019[76]). The Global E-Waste Monitor notes that “recycling activities are not keeping pace with the global growth of e-waste”; since 2010, the growth of e-waste generation “(has been) outpacing the formal collection and recycling by almost a factor of five” (Baldé et al., 2024[78]). Even countries in the European Union, after two decades of e-waste legislation and the “highest documented formal e-waste collection and recycling rate of 42.5%”, must increase collection rates further to meet EU targets (Forti et al., 2020[77]).
This objective of increasing formal e-waste collection and recycling is reflected in growing attention to reducing downstream environmental impacts of digital technologies. New policies are emerging to tackle e-waste and recycling. The EU WEEE Directive, for example, promotes re-use, recycling and other forms of recovery of WEEE. This aims to reduce the quantity of such waste to be disposed and to improve the environmental performance of the economic operators involved WEEE treatment (Directive 2012/19/EU, 2012) (EUR-Lex, 2012[81]).
Implementing circular economy solutions and “sustainability by design” could significantly decrease the e-waste related environmental impacts of digital technologies. Digital technologies, such as AI and IoT, can be applied to rethink product design and manufacturing, and extend the lifespan of products and their parts through predictive maintenance. They can also contribute to effective re-use of material through more efficient recycling and product recovery methods (One Planet Network, 2023[82]).
E-waste generated, and formally collected and recycled (kg per capita) in OECD countries, 2022
Notes: The indicator measures e-waste generated in a given year per inhabitant and the amount of e-waste that has been collected and recycled (not all e-waste that is collected is necessarily recycled). E-waste, as defined in the Global E-Waste Monitor, refers to all items of electrical and electronic equipment that have been discarded as waste without the intent of re-use. It includes cooling and freezing equipment, screens and monitors, lamps, large equipment (e.g. washing machines and solar panels), small equipment (e.g. vacuum cleaners, microwaves and electronic toys), and small IT and telecommunications equipment (e.g. mobile phones, personal computers and printers). The Global E-Waste Monitor estimates stocks of e-products for each country and the amounts being discarded in each year. Due to a lack of direct data on sales of e-products, new additions to the stock are estimated based on imports less exports. Domestic production is also included for EU countries and Norway. National authorities provide recycling and re-use figures to Eurostat under the Waste Electrical and Electronic Equipment (WEEE) Directive, based on surveys and administrative data from waste collectors and treatment facilities.
Source: The OECD Going Digital Toolkit, based on the Global E-waste Monitor and the OECD Annual National Accounts Database (OECD, 2019[76]; Forti et al., 2020[77]; Eurostat, 2023[83]), https://goingdigital.oecd.org/indicator/53.
Digital technologies are expected to significantly affect the transformation of several sectors central to achieving climate and economic resilience. Applying digital technologies to energy systems and networks, greening communication infrastructure and services, and improving the efficiency of the transport sector through digital adoption, offer great promise for accelerated climate action. Harnessing digital technologies to decarbonise these and other sectors will significantly increase humanity’s chance to reach net zero by 2050 (IEA, 2023[23]). Some scholars anticipate future years to bring a convergence of digitised broadband communication networks, electricity grids, the Internet, and mobility and logistics networks into increasingly integrated, digital and data-driven systems (Rifkin, 2022[84]).
Digital technologies have significant potential to bring positive sustainability impacts to nearly every sector of the economy. Key sectors that could become more sustainable from adoption of digital technologies include buildings and cities, heavy manufacturing like shipbuilding and steel production, farming and forestry, and green financing (Rolnick et al., 2022[85]). The OECD-FAO Agricultural Outlook 2021-2030 highlights that necessary improvements in productivity to feed the global population sustainably will not happen “without an important acceleration in digitalisation, technology, better data, and human capital” (OECD-FAO, 2021[86]).
Crucially, digital technologies are increasingly applied to climate predictions, forecasts and environmental modelling. Through its GreenData4All and Destination Earth projects, the European Union funds the development of a digital twin of the entire Earth system. It will be used to analyse the socio-economic impact of climate change and to develop strategies for climate mitigation and adaptation (Bauer et al., 2021[87]). While concrete evidence and macroeconomic projections have yet to materialise, digitalisation could contribute to “decouple economic activity from natural resource use and its environmental impacts”. This, in turn, would support the transition towards a resilient and circular global economy (Barteková and Börkey, 2022[88]).
A new green and digital energy paradigm: Digital technologies can enable the clean energy systems of the future
The creation of smart energy systems and networks is one of the most promising applications for digital technologies. In the wake of a global energy crisis with soaring electricity prices and mounting energy demands, the IEA anticipates a “historic turning point towards a cleaner and more secure energy system”. As clean technologies become cost competitive, a new clean energy paradigm is emerging in favour of a digitalised, decentralised and resilient clean energy grid (IEA, 2022[69]). Digital technologies play a fundamental role in enabling this transition. Technologies like AI, IoT and digital networks improve energy efficiency, reduce costs and accelerate clean technology innovation and diffusion across supply chains (IEA, 2023[23]).
The transformation of global energy systems from a centralised and fossil fuel-based system to a decentralised and renewable energy system requires significant flexibility. Such a system needs to integrate energy supply and demand from an increasing number of renewable energy sources as part of a resilient sustainable energy portfolio. Their flexible and decentralised nature make energy systems much more complex.
For these reasons, global energy systems will increasingly rely on digital technologies, especially those able to handle high levels of complexity and large amounts of data, to anticipate, manage and automate energy flows and prices. For example, smart energy storage can enable power-to-carrier (i.e. “power-to-X”) procedures, which refers to the conversion of surplus renewable electricity into various other carriers that can store energy, such as synthetic fuels or hydrogen (Lange and Santarius, 2020[89]).
Digital technologies can also support energy efficiency and innovation across clean energy supply chains themselves. They can do this, for instance, through innovation and clean technology discovery, development and deployment, natural resources materials extraction and processing, manufacturing and installation, operations, and end-of-life (IEA, 2023[23]).
According to the IEA’s 2023 Technology perspectives report, global energy flows are expected to become even more complex as more renewable energy sources come online. Electricity is set to become the largest energy vector, more than doubling in demand between 2021 and 2050 (IEA, 2023[23]). In addition, vast amounts of additional electrical loads are expected to be added to the electricity grid – from heat pumps to electric vehicles.
Tools like AI are optimal for analysing vast amounts of data to decipher patterns from complex data sets with many weights and parameters. They could play a significant role in managing and optimising future energy grids characterised by intermittent energy sources (IEA, 2023[23]). In a world where intermittent energy sources are predominant, connectivity is essential to co-ordinate the dispatch, transmission and distribution of electricity. As such, many of the environmental benefits of digital technologies and their underlying infrastructure derive from supporting the management of smart energy networks.
Deploying technologies such as fibre, 5G and AI systems can also optimise network management and reduce energy consumption. Applications such as “sleep mode” (enabled by machine learning) can help reduce energy consumption costs in mobile networks. Meanwhile, IoT and fibre-connected sensors can help optimise network energy management in buildings, cities and other critical infrastructure. Moreover, supporting the transition to fibre from legacy broadband access technologies could help achieve environmental sustainability goals for fixed broadband networks.
Re-imagining the electrical grid through a green lens
Several reports suggest that fibre-to-the-home (FTTH) networks may prove to be more energy efficient than traditional copper connections (OECD, 2022[90]) (Box 3.3). Around the world, new products such as AI-powered “home energy stations” monitor the energy demands of consumers and sell unused energy from devices like home solar panels back to the electricity grid. However, connectivity is essential for this type of AI and IoT applications to flourish. Connectivity divides, in particular in rural and remote areas, are a major challenge to overcome.
Other notable technology use cases for green energy grids include the deployment of smart meters in distribution networks. Smart meters can increase service quality and enable the introduction of innovative demand-side response measures by allowing customers to manage their energy consumption. Around 1.1 billion smart meters had been installed globally at the end of 2021 – already almost 40% of all residential meters (IEA, 2023[23]). Digital remote control and advanced protection devices can also manage bidirectional energy flows and identify grid faults quickly. Virtual Power Plants, for example, can integrate energy supply and demand, leveraging AI and IoT sensors (Nafkha-Tayari. et al., 2022[91]).
Further digital solutions include advanced voltage regulation at the distribution-grid level. Voltage regulation can increase the hosting capacity of the grid and enable integration of the increasing number of decentralised and intermittent sources of renewable electricity. In 2021, digital infrastructure accounted for 19% of global investment in electricity grids, with 75% of this amount in the distribution grid (IEA, 2023[23]). The increasing use of digital technologies in the energy sector can improve energy security through higher quality of supply and distributed energy sources. However, cybersecurity concerns pose a long-term risk for critical infrastructure such as power utilities (Casanovas and Nghiem, 1 August 2023[92]). Given digital security challenges, energy providers may prefer full-fibre connectivity for the energy grids.
AI-enabled digital twins of entire energy systems are a key tool that leverages digital technologies, combining clean technology use cases into a holistic and reliable solution for energy providers. As power systems need to increase flexibility by a factor of four by 2050, the digitalisation of power system management plays a key role in the net-zero transition (OECD, 2023[93]).
Digital twins, which represent a real-world system digitally by mirroring physical objects and processes, can improve forecasting, scheduling and control of power grids. They can also create advanced electricity systems to accommodate for flexible demand (Rolnick et al., 2022[85]). Digital twins exist for a variety of applications, such as Digital Twin Singapore for a multi-temporal digital virtual city (Singapore Land Authority, 2024[94]). They promote power systems that are closed-loop digital power grids, combining large amounts of data, machine learning, IoT and intelligent sensing for a national or even transnational digital twin (Bai and Wang, 2023[95]).
With respect to training AI systems, advances in data science can lead to fewer training runs involving smaller data sets and less complex models. In so doing, they can bring efficiencies more quickly than updating and modernising physical infrastructure such as data centres. For example, researchers at the Massachusetts Institute of Technology and at start-up MosaicML are training neural networks up to seven times faster by configuring AI algorithms to learn more efficiently (Leavitt, 18 July 2022[96]). With recent advances in large language models (LLMs), new generative AI tools have also emerged that can help advance understanding and accessibility of climate data (Box 3.2). The G7 Hiroshima Process on Generative AI launched in 2023 highlights generative AI’s potential role to help address pressing societal challenges, such as helping to solve the climate crisis and achieving the SDGs (OECD, 2023[97]).
Generative AI tools can help design greener products and make climate data more accessible
Generative AI systems are emerging as potential tools to support climate action across various domains. By leveraging advanced algorithms that learn from massive amounts of data, generative AI could be used to identify energy-saving options across sectors. This includes more efficient urban planning, greener product designs and manufacturing processes, better supply chain efficiency, and new ways of processing waste and optimising recycling. Large language models (LLMs) could also be leveraged to make climate data more accessible to a wider audience. For example, researchers trained an LLM on peer-reviewed climate papers and reports of the Intergovernmental Panel on Climate Change. In this way, they created an interactive chatbot that can make often complex climate science understandable and accessible to a wider audience (ChatClimate, 2023[98]). LLMs might also offer opportunities for climate and sustainability research. In biodiversity preservation, for example, generative AI has helped predict species coexistence patterns to promote biodiversity maintenance (Hirn et al., 2022[99]).
Further research is needed to assess in which areas of climate action LLMs could be most useful and to fully understand the environmental impacts of widely used generative AI tools. Some trends are already emerging. “Training” a machine-learning model (e.g. determining the weights, parameters and data to train a neural network, also referred to simply as “learning”) uses more energy than a single “inference” (e.g. using an AI chatbot to generate a response to a question). However, the inference stage overall typically is more energy- and water-intensive over the AI system’s life cycle. This is because such models are usually trained only a few times, whereas inference is executed repeatedly every time a system is used during its lifetime of deployment (OECD, 2023[58]).
The mass diffusion of generative AI tools across business and consumer products has placed heightened emphasis on the potential environmental impacts of “inference”, in addition to “training”. Initial research suggests that if generative AI systems include safeguards to manage energy use for training, they would provide net sustainability benefits to society (Larosa et al., 2023[100]).
Greening communication infrastructure and services is fundamental to sustainable and resilient digital transformation
Communication services and infrastructures have an impact on the climate both negatively (e.g. the high-energy consumption of data centres) and positively (e.g. through support for other parts of the economy). Apart from their direct impact on the environment, they have an indirect or catalyst effect on other sectors. According to the 2021 OECD Council Recommendation on Broadband Connectivity, the environmental sustainability of communication networks is critical for the future. It recommends minimising the negative environmental impacts of communication networks in two ways. First, policy makers should foster smart and sustainable networks and devices, such as the IoT. Second, they should encourage operators to periodically report on their environmental impacts and on the positive environmental effects of connectivity (OECD, 2021[101]).
The Recommendation was a building block to the “G20 Guidelines for Financing and Fostering High-Quality Broadband Connectivity for a Digital World”, developed with the support of the OECD (G20, 2021[102]). The G20 Guidelines recommend incentivising “communication network operators and other sectors, such as the transportation and energy sectors, to co-operate in network development and financing activities in order to minimise costs, disruption and environmental impacts” (G20, 2021[102]).
In recent years, the communication infrastructure and services industry has promoted the sustainability of communication networks to reach net-zero targets (Box 3.3). According to the IEA, several large network operators have improved their networks’ energy efficiency through innovative technologies, significantly reducing energy use. For example, despite a growing demand for energy, the communication operator company Sprint reduced the energy intensity of its network by more than 80% between 2014 and 2019, keeping its total network energy consumption flat (IEA, 2023[17]).
Beyond “future proofing” aspects of symmetrical broadband speeds and the scalability of networks, the transition to fibre can also promote environmental sustainability. Several reports suggest that FTTH networks may prove to be more energy efficient than traditional copper connections (OECD, 2022[90]). According to one report, fixed fibre networks consumed on average 0.5 Watts (W) per line (Arcep, 2019[103]). This translates into three times less energy consumption than an ADSL line (1.8 W) and four times less than a traditional Publicly Switched Telephone Network line (2.1 W). Another study found that energy efficiency gains achieved from 5G deployment will begin in 2023 and be clear by 2028 in the most densely populated areas. However, it will be far more modest in more sparsely populated areas (Arcep, 2022[104]).*
In recent years, the communication industry has undertaken various efforts to promote the sustainability of networks. Three large operators in Europe, for example, have categorised fibre rollout as part of their environmental sustainability agenda. They have linked “green” credit funding to achieve this objective. KPN, a fixed and mobile operator in the Netherlands, refinanced its credit line by tying the new interest rates to the company’s performance in its sustainability strategy, such as fibre deployment and reduction of energy consumption (Lenninghan, 2021[105]). KPN plans to invest EUR 3.5 billion (USD 3.99 billion)** by 2024 as it aims for nationwide fibre deployment (Telecom Review, 2020[106]). In a similar way, the Swedish operator Telia used two “Green Bonds” funds for fibre investments. This responded to the company’s vision of fibre rollout as energy saving and a key enabler of IoT solutions that help reduce carbon emissions (Lenninghan, 2021[105]). For example, fibre-connected street furniture may enable IoT sensors across cities to optimise energy consumption and traffic management, resulting in fewer CO2 emissions. Meanwhile, Telefónica issued its first “sustainable perpetual hybrid” bond amounting to EUR 1 billion (USD 1.142 billion)** in February 2021. This will finance environmental projects in Spain, Germany and Brazil, focusing on the transformation of copper networks to more reliable and energy-efficient fibre (i.e. 85% more energy efficient) (Telefónica, 2021[107]).
A WIK report assessed the environmental effects of changes in the fixed broadband technology mix in Europe. Assuming that power sources remain unchanged, a migration from fixed broadband technology in the European Union to 100% fibre would reduce CO2 emissions from 15.5 million tonnes to 3.2 million tonnes per year (i.e. a 79% yearly reduction) as FTTH is more energy efficient (WIK-Consult, 2020[108]).
* The study identifies two scenarios based on identical traffic growth: a 4G-only network and a network that combines 4G and a 5G deployment. Initially, 5G will generate an increase in energy consumption – for a length of time that depends on different 5G rollout scenarios. After which 5G deployment will enable total energy savings of up to ten times 2020 consumption levels by 2028, compared to a scenario of 4G-only network densification, as well as a corresponding decrease of greenhouse gas (GHG) emissions of up to eight times 2020 GHG emissions. In less densely populated areas, however, where traffic density is lower, virtually non-existent gains will be seen until 2025 at the earliest, and by 2028 at the latest.
** An exchange rate of EUR 0.876/USD for the year 2020 from OECD.stat has been used.
Source: OECD (2022[90]).
Many communication regulators across OECD countries actively support decarbonisation of the sector. They do this either through their mandates or through inter-agency co-operation to achieve digital policy objectives that require a whole-of-government approach (OECD, 2022[109]). For example, the government of France tasked the communication regulator, Arcep, and the agency for ecological transition (ADEME), to quantify the current and future environmental footprints of digital technologies. In 2022, the two agencies assessed the current impact of ICTs on the environment in two volumes (Arcep, 2022[110]). In 2023, the third volume provided a forward-looking assessment (2030-50) (Arcep, 2023[111]). Moreover, compared to 2021 (OECD, 2022[109]), partial or full responsibilities of communication regulators in 2023 increased notably with regards to the environmental sustainability of networks. In 2023, 52.5% of OECD communication regulators reported responsibilities in this area (see Chapter 2).
Environmentally responsible practices and objectives of communication networks include the following:
Reduction of energy consumption of network operations and the usage of renewable energy sources;
Reduction of pollution, radiation and other hazards of networks;
Adoption of environmentally responsible policies for network construction such as land-use policies, cell tower construction and data processing centres;
Reduction of environmental impacts of electronic equipment and terminals once discarded e-waste, by adopting proper recycling and safe disposal practices; and
Creation of more sustainable products, using a minimum of hazardous materials and allowing for longer useful lives rather than planned obsolescence (OECD, 2022[90]).
Moreover, as explored in the DEO Volume 1 Spotlight “Next generation networks and the evolving connectivity ecosystem” (OECD, 2024[13]), policy makers are considering environmental sustainability as a key value for 6G technologies and use cases for the next decade. Some stakeholders even call it the “green G”.
The large enabling effects of communication and broadband for climate action and environmental stewardship are equally important to minimise the negative environmental impacts of communication networks. Broadband is often regarded as the foundation of the SDGs. For instance, connectivity enables the IoT across sectors such as energy, transport and agriculture. Massive machine-to-machine communication services, a subset of IoT, comprise the vast amount of sensors used in cities (e.g. electrical grids and highways), in industry (e.g. sensors within machines) and in the agricultural sector (e.g. sensors measuring humidity levels to improve water efficiency or better predict crop yields), among others (OECD, 2018[12]).
Enabling adoption of smart devices, such as IoT, can have a positive impact on the environment through a wide range of applications. Such applications range from smart electrical grids, fleet automation and precision agriculture to predictive maintenance, connected forests and traffic management systems that reduce transport congestion in smart cities. Smart grids are one application being fostered in many countries. In 2019, the Irish communication regulator (ComReg) assigned its 400 megahertz (MHz) Band Spectrum Award for the use of smart grids. The award complemented the Irish government’s climate policies, with smart grids described as “an efficient utility network system typically using digital automation technology for monitoring, control, and analysis within the supply chain” and a key enabler for the reduction of GHG emissions (OECD, 2018[12]). For its part, Germany awarded an exclusive licence in the 450 gigahertz (GHz) band for a smart electrical grid private network in February 2022 to “450connect”, a consortium of German regional and municipal energy and water utilities along with energy companies (Jones, 2022[112]).
Communication policy and regulation play a key role with regards to Earth observation, which is important to support climate mitigation. For example, spectrum policy helps enable Earth observation satellites that support several use cases in agriculture, as well as disaster preparedness, and weather and climate monitoring. Data from such satellites are expected to play an even greater role as countries grapple with the impact of climate change (OECD, 2022[113]). In addition, the International Telecommunication Union is discussing the role that “smart” submarine cables (i.e. equipped with scientific sensors) could play in providing real-time data for ocean climate monitoring and disaster mitigation (e.g. tsunamis) (ITU-T, 2023[114]).
As a significant contributor to GHG emissions, global transport systems stand to benefit greatly from energy efficiency gains enabled by digital technologies
With demand for travel increasing, the global transportation sector has made only modest progress at lowering emissions. The sector, excluding the manufacturing of motor vehicles and other transport equipment, was responsible for around a quarter of global energy sector GHG emissions in 2022 (IEA, 2023[60]).
Following the COVID-19 pandemic, demand for passenger road transport, freight transport (road, shipping and air), and commercial aviation have rebounded. Consequently, the sector has not significantly reduced its emissions. Moreover, demand for transport is expected to increase by 2050. Further concerns arise from unregulated and often uncontrollable GHG emissions like methane and nitrous oxide, and a shift away from low-carbon collective transportation systems like public buses to individual mobility solutions (EEA, 2022[115]).
Digital technologies can help reduce the environmental impact of global transportation networks. Digital technologies can reduce overall demand for travel and transportation through, for example, videoconferencing and teleworking. They can also help increase fuel efficiency and infrastructure longevity through AI-enabled digital twins to forecast energy needs, and IoT sensors for predictive maintenance (EEA, 2022[115]).
Shared mobility such as on-demand ride services or vehicle sharing can also reduce overall passenger transport activity. Freight routing and consolidation, such as smart shipment bundling, can significantly reduce freight trips (Rolnick et al., 2022[85]). The International Transport Forum describes how smart transport systems can improve operational efficiency of non-urban freight movements such as long-haul trucking, decreasing GHG gas emissions and costs through increased asset sharing and the use of high-capacity vehicles (ITF, 2023[116]).
Global transportation networks can benefit from digital technologies
The carbon intensity of transportation can be further lowered through digital technologies that improve the performance of low-emission vehicles and batteries, resulting in less overall demand for electricity and critical minerals. They can also speed up clean technology discovery and deployment by accelerating the R&D of alternative fuels such as synthetic fuels or hydrogen. Such alternative fuels hold great potential for decarbonising sectors that are difficult to electrify, including aviation, long-distance trucking and maritime shipping (Lange and Santarius, 2020[89]).
Connected and fully automated vehicles may help reduce air pollution but are also expected to generate negative impacts on the environment. These vehicles, commonly called “self-driving” cars or “autonomous vehicles” (OECD/ITF, 2015[117]; OECD, 2018[12]), include truck platooning (i.e. linking of two or more trucks in convoy using connectivity). They may help reduce road congestion, and hence air pollution, by traffic management techniques or simply by wind breaking (in the case of platooning). This is particularly important due to the contribution of road congestion to GHG emissions. At the same time, fully automated vehicles, using advanced wireless networks, are expected to produce massive amounts of data (OECD, 2018[12]), which have their own environmental footprint.
Despite breakthroughs in efficiency for maritime transport and aviation, more progress is needed
Despite large efficiency increases in global transportation in the past decades, there is still huge potential for optimisations in sectors such as maritime transport and aviation. A research consortium from Google Research, American Airlines and Breakthrough Energy, for example, has demonstrated how to reduce the contrails of airplanes by up to 54% in test flights using AI to develop contrail forecast maps. Contrails, which are clouds formed when water freezes around aerosols in airplane exhausts, generate around 35% of the global warming impact of aviation.
Given these trends, AI-based predictions could significantly reduce GHG emissions from aviation in a highly cost-effective manner (Elkin and Sanekommu, 2023[118]). Another example is RASMUS, which combines AI with oceanographic models to calculate shipping routes that leverage small dynamic ocean currents and swirls. The optimised routes could result in GHG emissions savings of up to 10% for shipping operators (Christian-Albrechts-Universität zu Kiel, 2023[119]).
Many experts agree that a narrow focus on increasing the efficiency of the transport sector will be insufficient to meet the sector’s climate goals because of rebound effects. The adoption of digital technologies in transportation has significantly increased energy efficiency. However, in many cases they have not decreased overall emissions as efficiency gains were offset by an increase in transportation demand (Creutzig et al., 2015[120]; Lange and Santarius, 2020[89]).
Analysts have raised similar concerns about the development and deployment of fully automated vehicles. Such vehicles offer the prospect of reducing fuel consumption and increasing vehicle occupancy rates. However, any gains could be offset by an increase in overall vehicle usage and road traffic as vehicle transportation becomes even more accessible (Barcham, 2014[121]; Rolnick et al., 2022[85]; Silva et al., 2022[122]).
Experts also question the energy requirements stemming from the IoT, data sharing and computing needs demanded by fully automated vehicles to deploy accurately and safely. One study estimates that emissions produced from a global fleet of fully automated vehicles would match that of all global data centres today in a 95% adoption rate scenario (Sudhakar, Sze and Karaman, 2022[123]).
Policy makers could encourage a shift to more efficient travel modes to decarbonise transportation
One of the most important ways to decarbonise transportation systems is called “modal shift”, which describes strategies to incentivise the shift from carbon-intensive to low-carbon modes of transportation. The International Transport Forum encourages policy makers to enable modal shift and demand management where they are most effective. This is typically in urban environments and short-distance intercity and international travel (ITF, 2023[116]). Digital technologies can enable such modal shift through more efficient transport planning options. For example, they can nudge passengers towards low-carbon transport options, and enable low-carbon transport modes like vehicle sharing (Rolnick et al., 2022[85]).
Future transport systems could make use of multimodal digital mobility services, enabled by technologies like AI, cloud computing, mixed reality and predictive analytics. Such systems can also integrate road, rail, water, and air transport at urban, interurban and rural scales. They help passengers compare travel options and facilitate access to low-carbon, multimodal modes of public transport. In so doing, they enable transport planners to design a public transportation network that is both environmentally beneficial, reliable and efficient, and highly attractive for passengers (EEA, 2022[115]).
Digital technologies drive greener consumption choices in economies and societies
Digital technologies can be used to nudge consumers towards greener choices (Sunstein and Reisch, 2013[124]) – from reducing energy consumption (OECD, 2017[125]; Rivers, 2018[126]) to more sustainable online shopping (Banerjee et al., 2022[127]; Michels et al., 2022[128]). A nudge is generally “any aspect of the choice architecture that alters people’s behaviour predictably without forbidding any option or significantly changing their economic incentives” (Thaler and Sunstein, 2009[129]). Nudges include data and notifications about energy and other consumption and proposing options for more sustainable alternatives for online purchases.
Such digital nudges could play a significant role in the green transition given that household spending accounts for around 60% of gross domestic product across the OECD (OECD, 2024[130]). At the same time, the potential positive effects from nudging consumers towards greener choices can be difficult to quantify. This further underlines the need for data collection and net benefit analyses of digital technologies on the environment (République Française, 2021[131]).
Supported by digital technologies, the trend towards teleworking may support the green transition. The COVID-19 pandemic saw an acceleration of teleworking enabled by digital technology. This allowed many businesses to continue operating with personnel working from home and using tools such as videoconferencing, cloud services and virtual private networks to communicate and work together. Early evidence suggests that around one-third of employed persons would like to continue being able to telework. A higher share of businesses also anticipates offering teleworking than before the pandemic – with a greater proportion of employees expected to make use of the option to telework (Ker, Montagnier and Spiezia, 2021[132]).
At the same time, digital technologies can be used in a similar way to promote interests that run counter to environmental sustainability goals. For instance, through manipulatory nudging, choice architects can influence consumption choices towards less sustainable alternatives using the same methods described for more greener choices (Sunstein and Reisch, 2013[124]).
Figure 3.1 shows that digital technologies can have negative environmental impacts through the direct technological life cycle, as well as through enabling and systemic effects. This includes both intended and unintended consequences of the application of digital technologies, such as the acceleration of emissions-intensive activities through AI. It also comprises system-level impacts like rebound effects in autonomous driving or negative consumption lifestyle changes through digital advertising (Kaack et al., 2022[61]). This is why some scholars have proposed the concept of “digital sobriety” or “digital sufficiency”, which calls for a new “Digital Green Deal” that puts digital technologies in the service of a deep sustainability transformation (D4S, 2022[133]).
Policy responses acknowledge the link between the green and digital twin transitions
Policy makers increasingly consider both digitalisation and environmental sustainability on their agendas. They may include them, for example, in national strategies and recovery plans; in dedicated strategies for digital technologies and the environment; or through standalone national strategies for digitalisation or for the environment. Elevating these priorities to a strategic level while adopting a whole-of-government perspective that integrates digital and environmental policies across domains contributes to their effectiveness (OECD, forthcoming[16]).
Recent declarations and commitments at the OECD and beyond also emphasise the importance and opportunity of leveraging digital technologies for climate and economic resilience. This section provides a selection of policy responses taken by countries broadly grouped along three notable trends: i) aligning the vision, values and objectives of the green and digital twin transitions; ii) measuring, minimising and mitigating negative environmental impacts of digital technologies; and iii) accelerating innovation and adoption of green digital technologies solutions.
Aligning the vision, values and objectives of the green and digital twin transitions
Although the green and digital twin transitions have emerged together as policy priorities, such transitions are not always automatically aligned. Digital technologies are often designed and deployed to achieve social and economic gains first and foremost. They do not necessarily consider whether such technologies are “sustainable by design” and advance sustainability goals.
In recent years, policy makers have begun to align the vision, values and objectives of the green and digital twin transitions. Policies fit for the future will no longer consider only how to achieve economic productivity gains; policies become inseparable and synonymous with actions that also protect the planet. Designing such policies starts with the fundamental alignment of vision, values and objectives.
Many key intergovernmental organisations recognise the importance of vision alignment:
The OECD Recommendation on Information and Communication Technologies and the Environment (OECD 2010[4]), adopted over a decade ago, was pioneering in recognising the connection between digital technology and environmental sustainability. Its ten principles helped lay the groundwork for using ICTs to improve environmental performance, increase energy efficiency and combat climate change.
The UNEP-led CODES network calls for aligning “the global practice and discourse around digital advancement with sustainable development”. It introduced several initiatives around this goal, including a World Commission on Sustainability in the Digital Age to streamline the twin transitions across the United Nations and beyond (CODES, 2022[10]). CODES also stresses the importance of considering both Indigenous and modern understandings of sustainability and protecting Indigenous land and data rights (CODES, 2022[10]).
The Institute of Electrical and Electronics Engineers underscores the need to align the development and use of technology with ethical and environmentally responsible practices in its principles for Strong Sustainability by Design (IEEE SA, 2023[134]).
The European Union Recovery and Resilience Regulation and its implementation across EU member states offers another example of an integrated approach that accounts for digital and environmental priorities as part of overall post-COVID growth and recovery plans (European Union, 2021[135]). Research from the European Commission also emphasises the importance of aligning the twin transitions to ensure the term not only describes parallel green and digital transitions, but also aims to unite both into one (Muench et al., 2022[11]).
Some countries have recently adopted national plans targeting the digital-environment nexus. In 2023, the Korean government established a plan to promote carbon neutrality through digital transformation. The initiative aims to secure green digital transformation technologies and infrastructure, and to disseminate them for use across the public and private sectors. Specifically, the initiatives cover developing technologies to reduce carbon emissions by sector, developing and applying low-power and high-performance data centre technologies, developing core technologies for low-power networks, laying the foundation for using carbon-neutral data and establishing a carbon-neutral decision-making support system (Government of Korea, 2023[136]).
Similarly, Finland, France and Germany have adopted overarching national plans for digital technologies and the environment. These seek to leverage the potential of digitalisation for climate objectives, while ensuring a sustainable digital transformation (Germany Federal Ministry for the Environment, Nature Conservation and Nuclear Safety, 2020[137]; Finland Ministry of Transport and Communications, 2021[138]; Government of France, 2021[139]):
France’s Law No. 1485 of 2021 on reducing the digital footprint on the environment recognises the impact of digital technologies in an integrated approach to environmental sustainability, cutting across policy areas. The law introduced and modified specific legal provisions in a broad array of legislation, such as consumer, environmental, electronic communications, intellectual property and commercial law (République Française, 2021[131]).
Germany’s Digital Policy Agenda for Environment covers four types of measures, including measures both to reduce resource use by digital technologies and those targeted at using AI to tackle environmental challenges. Germany has also proposed the Energy Efficiency Act would be applicable to data centres. If implemented, it would require efficiency benchmarks and re-use of excess heat (Bundesregierung, 2023[140]).
Promoting policies to measure, minimise and mitigate negative environmental impacts of digital technologies
Ensuring that digital transformation is aligned with efforts to minimise the environmental impacts of digital technologies is at the core of country policies seeking to promote an environmentally sustainable future. Such policies help not only to align digital transformation with climate targets, but also to increase the benefits from their application across sectors.
Countries have started to introduce measures to gauge, minimise and mitigate the environmental impacts of digital technologies. These include monitoring and tracking of impacts, sustainable procurement of digital infrastructure and the introduction of digital product passports (Muench et al., 2022[11]). While efficiency improvements, durability and recycling will play a critical role, some researchers and government agencies have called for the principle of “digital sufficiency”. This entails only using digital technologies where they are necessary and offer a clear advantage over low-tech or no-tech solutions (D4S, 2022[133]).
In many cases, initiatives focus on the end-of-life phase and extending the lifespan of digital products. France’s Law No. 1485 of 2021 seeks to reduce the environmental footprint of digital technology through several measures. These include prohibiting the planned obsolescence of devices; requiring the distribution of information to consumers about how to optimise equipment performance to expand its lifespan; and promoting annual national campaigns for returning electronic equipment (République Française, 2021[131]). The German Digital Policy Agenda for the Environment advocates for binding requirements on hardware manufacturers to expand product lifespans. It also highlights the responsibility of both the private and public sectors in reducing e-waste (Germany Federal Ministry for the Environment, Nature Conservation and Nuclear Safety, 2020[137]). The United Kingdom’s “Greening Government: ICT and Digital Services Strategy 2020-2025” similarly focuses on the end-of-life phase, seeking to “map and account for ICT at end of life” to increase transparency, as part of procurement processes, and to advance a circular economy (Defra, 2020[141]).
Policy makers are analysing indirect impacts from new technologies on the environment
In addition to life-cycle approaches, policy makers are factoring indirect environmental impacts into policy design, such as enabling and systemic effects (indirect impacts). Enabling and systemic effects may arise from the use of new technologies themselves or be scaled by policies aimed at promoting them. With rebound effects, policies may have unintended outcomes, such as increasing consumption by making products and services more efficient. Rebound effects may not be easily foreseeable or identifiable due to the complexity of environmental systems, value chains and social behaviour involved in the assessment of their impact.
At the national level, policy makers seek to address the complexity of indirect environmental impacts in dedicated strategies for digital policy for the environment. For example, both Finland and Germany underscore the importance of understanding and measuring rebound effects of digital solutions to environmental challenges (Germany Federal Ministry for the Environment, Nature Conservation and Nuclear Safety, 2020[137]; Finland Ministry of Transport and Communications, 2021[138]). A French observatory established in 2021 aims to quantify and analyse direct and indirect environmental impacts of digital technologies, notably AI (République Française, 2021[131]). In France, the communication regulator, Arcep, is conducting environmental assessments of the ICT sector. In 2021, the government of France strengthened Arcep’s powers to provide clear environmental impact information-gathering authority. This covers not just network operators, but also online communication service providers, data centre operators, consumer device manufacturers, network equipment suppliers and operating system providers (République Française, 2021[142]).
At the European level, the proposed European Union AI Act would require high-risk AI systems to be designed to enable “the recording of energy consumption, the measurement or calculation of resource use and environmental impact”. The draft AI Act also highlights the potential contribution of AI systems to environmental monitoring; conservation and restoration of biodiversity and ecosystems; and climate change mitigation and adaptation (European Parliament, 2023[143]).
Measuring the environmental impact of digital technologies remains a challenge
Although several policies target measurement issues, regarding both the impact of digital technologies on the environment and of related policies, this remains a challenging area due to the myriad of frameworks and factors involved (OECD, forthcoming[16]). According to OECD (2010[4]), governments are encouraged to measure the impact of digital technologies themselves on the environment (through comprehensive and comparable metrics) and to measure the impact-related policies. In both cases, digital technologies play a role in advancing the ability to track and measure impact. Moreover, in both cases, the OECD can support countries in developing and applying co-ordinated and comparable measurement frameworks.
There is a need to further explore circular economy models for digital transformation with an emphasis on the entire life cycle to assess the environmental impact of networks and devices (OECD, 2022[113]). Regarding mobile handset acquisition models, previous OECD work underscores the need to properly reflect the societal cost of extraction and waste disposal of the metals used in mobile handsets (OECD, 2013[144]).
The OECD report “Case Study on Critical Metals in Mobile Devices” provides valuable recommendations for the management of critical metals in mobile phones through their life cycle (OECD, 2012[145]).11 Access to reliable data and harmonised methodologies is a prerequisite for the pursuit of the objectives identified. In many OECD countries, regulators and/or ministries lack a mandate to collect data on the environmental impact of ICTs.
Accelerating innovation and adoption of green digital technology solutions
Digital technology and innovation are major building blocks to reach environmental goals and achieve the deep cuts in emissions needed to transition to a net-zero carbon world. Innovation is crucial because it can help reach environmental sustainability objectives, and is also the main source of modern economic growth. This implies that technology and innovation may help enable a green, more resilient future that goes in hand in hand with new growth opportunities and strengthened productivity growth. However, after rapid progress in the early 2000s, low-carbon innovation efforts (e.g. as measured by patent filings and public spending on energy R&D) started to decline around 2012. This decline occurred despite the ambitious climate objectives set out in the 2015 Paris Agreement (Cervantes et al., 2023[146]).
Technological progress fuelled by public and private investments can reduce the costs of emissions reduction policies. This is demonstrated by sharp declines in the costs of batteries and solar, which have both experienced a 90% reduction over the past decade (Cervantes et al., 2023[146]). However, reaching net zero by 2050 requires not only the rapid deployment of currently available technologies. It also demands further innovation in breakthrough technologies not yet on the market (Cervantes et al., 2023[146]). Strengthening innovation, along with technology diffusion, around the green and digital twin transitions is therefore essential to reach carbon neutrality and other environmental goals. Much of this innovation to increase efficiency and resource productivity relates to adoption of digital technologies, as seen in previous sections.
Several jurisdictions are investing in innovation to advance research and development of digital technologies
At least half of global reductions in energy-related GHG emissions through 2050 will rely on technologies not yet fully available for commercial use as they are at the demonstration or prototype phase (IEA, 2021[147]). Costs, access and availability of technology required for such solutions all hamper the innovation and commercialisation of such technologies. To address these challenges, governments have incorporated a mission-oriented approach to drive technological breakthroughs (OECD, 2023[93]; OECD, 2023[148]).
Governments are stepping up direct investments and creating incentives for other stakeholders to do the same, to advance R&D and innovation in digital technologies in support of the transition to a green and circular economy. Denmark provides funding, subsidies and tax deductibility, among other measures, for green research on technologies to capture and store CO2 (Ministry of Finance, 2021[149]). Finland invests in emerging technology for clean energy production, including by using AI to reduce energy consumption and emissions (Finnish Government, 2021[150]). Mexico’s National Institute of Ecology and Climate Change oversees the co-ordination of technological and scientific research and projects in co-operation with research institutes (Government of Mexico, 2022[151]).
Several jurisdictions are strengthening the underlying role of data access to enable digital innovation. The “All Data 4 Green Deal”, a consortium of 12 partners jointly funded by Switzerland, the United Kingdom and the European Union, seeks to co-design a “Green Deal” common data space. This will enable the interoperable combination and integration of data from a variety of sources to support innovation; access; and informed decision making related to climate change, pollution and biodiversity (AD4GD, 2023[152]). For its part, Austria’s Mobility plan refers to the role of data to support better decision making and innovative, energy-efficient and sustainable solutions in the transportation sector (BMK, 2021[153]).
Countries are also implementing policy measures to develop the skills needed for harnessing digital technologies for environmental sustainability, while reducing negative impacts. For both the earliest school age and university entry level, France developed training in “digital sobriety” and the impact of digitalisation on the environment. This is one of several measures to reduce the environmental footprint of digital technologies (INSP, 2023[154]). In Switzerland, the Digital Strategy seeks to embed environmental concerns into the development of digital skills (Federal Chancellery of Switzerland, 2023[155]). For their part, Finland and Germany seek to include “green coding”, or environmentally sound software design, as part of the training of computer programmers (Germany Federal Ministry for the Environment, Nature Conservation and Nuclear Safety, 2020[137]; Finland Ministry of Transport and Communications, 2021[138]).
The European Commission regards digital technologies as a cornerstone of the European Green Deal to make the European Union climate neutral by 2050. It expects technologies like AI and IoT to enable the green transition in the agricultural sector, buildings and construction, electricity systems, energy-intensive industries, and transport and mobility (Muench et al., 2022[11]). Through the NextGenerationEU fund, the European Union has committed over USD 842 billion (EUR 800 billion)12 to build a “greener, more digital and more resilient future”. By 2050, the European Union plans to be the first climate-neutral society, while embracing technology (European Commission, 2023[156]).
The United States has passed three laws in recent years to support connectivity, digital technologies and initiatives related to climate change. The Inflation Reduction Act (IRA), signed in August 2022, allocated USD 500 billion in funds and tax breaks for clean energy and climate resilience investment programmes in various sectors (Badlam et al., 2022[157]; White House, 2022[158]). The Infrastructure Investment and Jobs Act of 2021 allocated USD 65 billion to broadband infrastructure. Finally, the “CHIPS and Science Act” was passed in 2022. Together, the three laws commit USD 2 trillion in public funds to connectivity, digital technologies and the fight against climate change (Badlam et al., 2022[157]). The IRA incentivises private companies, including those working on ICTs, to invest in clean production and development practices, as well as to develop talent in the field of clean technology (Badlam et al., 2022[157]).
As digital technology development and innovation progress, some potential policy questions arise. Like any product or service, digital technologies have environmental footprints. The challenge is to balance between minimising such footprints and fostering the positive, enabling environmental impacts of digital technologies throughout economic sectors. Examples of policy questions at the intersection of digital technologies and the environment include the ones below.
Are local energy grids ready to support the green and digital twin transitions? As the world becomes increasingly digitalised, data centres and data transmission networks are emerging as an important source of energy demand. However, at a local and regional level, energy grids may have limited capacity to support future levels of digital adoption. Already, data centre providers face strained national power supplies and rising energy costs in some locations, depending on the local energy grid mix and grid capacity. The energy demands of large data centres have led some jurisdictions to consider or impose moratoria and zoning rules on new data centre construction. This aims to ensure sufficient energy is available for other purposes like residential housing. In designing policies for the energy grids of the future, further understanding is needed around how changing demand for digital technologies translates into overall energy demand (IEA, 2023[17]).
Will computing power need to be managed as a national resource, with data centres becoming a new utility? Countries have acknowledged the increasing importance of computing power to enable broad digital transformation throughout economies and to train newer innovations like frontier AI models. Computing power is increasingly being viewed as a national resource to be carefully managed through, for example, hardware trade restrictions. As the critical infrastructure behind digital transformation, data centres are set to play a key role in enabling future productivity gains from digital technologies. However, data centre energy use is placing pressure on energy grids and becoming a key driver of rising costs. Policy makers need to decide if data centres should be treated and regulated like utilities.
A window of opportunity: Policy alignment for a green and digital future
Policy alignment and harnessing the potential of the digital and green twin transitions are essential to securing a future that is innovative, inclusive and sustainable. With the timeframe narrowing to avoid the most catastrophic impacts of a changing climate, leaders across all stakeholder groups and countries must share policy good practices and work together in support of a resilient future.
Promoting national policies in support of the twin transitions. Environmental considerations have gained increasing importance on policy agendas globally. Many consider climate change as the major challenge in the years ahead. Countries have launched and implemented significant roadmaps, policies and legislation to fight climate change and preserve biodiversity across the United Nations, the OECD and on a national level. Many economic recovery packages emphasise structural reforms to reduce emissions by acknowledging that “digital” and “green” policies are intertwined, and together can achieve sustainable growth (OECD, 2022[90]). The concept of the twin transitions is a key framework for policy makers to understand as countries race to achieve their sustainability targets in the crucial few decades ahead (Muench et al., 2022[11]).
Standardising the measurement of environmental impacts of digital technologies. Measurement of the environmental impacts of digital technologies is limited by a lack of common terminology, recognised standards, and varying or optional reporting requirements. Specific standards and policies are underdeveloped compared to other environmental, social and governance reporting requirements (OECD, 2022[20]). Harmonised indicators across countries will need to reflect a holistic understanding of the environmental impacts of digital technologies throughout their life cycle and applications. For example, the IEA recommends data centre providers collect and report sustainability data beyond energy consumption to include “embodied” life-cycle emissions from raw material extraction or end-of-life-disposal (IEA, 2023[17]). Focusing on select indicators could have unintended consequences and the compliance of such metrics could also be considered. Moving beyond measurement, the concept of “sustainability by design” calls for embedding sustainability standards and practices into the design of technological solutions from the beginning (IEEE SA, 2023[134]). Computer scientists and engineers can also work to better understand how their products and services generate various sustainability impacts in the real world (CODES, 2022[10]).
Facilitating intergovernmental co-operation to achieve climate targets. OECD legal instruments have increasingly recognised the interlinkages between digital technologies and the environment. The 2010 OECD Recommendation on Information and Communication Technologies and the Environment encourages the development of comparable indicators for the environmental impacts of ICT goods, services and applications. The 2019 OECD Recommendation on Artificial Intelligence, updated in 2024, underlines that AI systems should support beneficial outcomes for people and the planet and explicitly references environmental sustainability as a key concern. OECD countries also consider it critical to analyse the environmental impact and sustainability of communication networks. This is demonstrated by the 2021 OECD Recommendation on Broadband Connectivity, which stresses the need to minimise the negative environmental impacts of communication networks. The OECD also has an ongoing horizontal project on Net Zero+ policies to achieve climate and economic resilience in a changing world (OECD, 2023[93]), as well as a Horizontal Project on Going Digital Phase IV with a pillar focusing on “twin transitions”.
The digital transformation of global economies and societies is accelerating at a rapid pace with technologies like AI, broadband and IoT shaping how people live, work and think, bringing productivity gains and improving lives. At the same time, human activity is profoundly transforming the planet, with the climate crisis endangering the natural foundations on which humanity depends. This transformation has been profound and persistent, leading scientists to make ever starker warnings. In this complex interplay between technology and sustainability, the future lies in humanity’s ability to promote innovation that aligns the digital and green twin transitions, and in crafting policies that usher in a world where digital technologies not only bring economic gains but also preserve and protect the planet.
References
[152] AD4GD (2023), All Data 4 Green Deal - An Integrated, FAIR Approach for the Common European Data Space, website, https://ad4gd.eu (accessed on 19 January 2024).
[67] Andrews, D. et al. (2021), “A circular economy for the data centre industry: Using design methods to address the challenge of whole system sustainability in a unique industrial sector”, Sustainability, Vol. 13/11, p. 6319, https://doi.org/10.3390/su13116319.
[111] Arcep (2023), “Christophe Bechu, Agnès Pannier-Runacher and Jean-Noël Barrot receive the forward-looking assessment of the digital environmental footprint in 2030 and 2050, from ADEME and ARCEP”, 6 March, Press Release, Arcep, Paris, https://en.arcep.fr/news/press-releases/view/n/environment-060323.html.
[110] Arcep (2022), “The digital environmental footprint in France: ADEME and Arcep submit their first report to the government”, 19 January, Press Release, Arcep, Paris, https://en.arcep.fr/news/press-releases/view/n/the-environment-190122.html.
[104] Arcep (2022), “The Environment: 5G and networks’ environmental footprint: Arcep engages in new work to inform the debate and identify levers of action”, 14 January, Press Release, Arcep, Paris, https://en.arcep.fr/news/press-releases/view/n/the-environment-140122.html.
[103] Arcep (2019), “Réseaux du futur: L’empreinte carbone du numérique”, Note, No. 5, Arcep, France, https://www.arcep.fr/uploads/tx_gspublication/reseaux-du-futur-empreinte-carbone-numerique-juillet2019.pdf.
[157] Badlam, J. et al. (2022), “McKinsey: The Inflation Reduction Act: Here’s what’s in it”, 24 October, McKinsey & Company, https://www.mckinsey.com/industries/public-sector/our-insights/the-inflation-reduction-act-heres-whats-in-it.
[95] Bai, H. and Y. Wang (2023), “Digital power grid based on digital twin: Definition, structure and key technologies”, Energy Reports, Vol. 8/4, pp. 390-397, https://doi.org/10.1016/j.egyr.2022.10.328.
[78] Baldé, C. et al. (2024), Global E-waste Monitor 2024, International Telecommunication Union and United Nations Institute for Training and Research, Geneva/Bonn, https://ewastemonitor.info/wp-content/uploads/2024/03/GEM_2024_18-03_web_page_per_page_web.pdf.
[127] Banerjee, S. et al. (2022), “What works best in promoting climate citizenship? A randomised, systematic evaluation of nudge, think, boost and nudge+”, Research Square, https://doi.org/10.21203/rs.3.rs-1534312/v1.
[121] Barcham, R. (2014), Climate and Energy Impacts of Automated Vehicles, Goldman School of Public Policy, University of California.
[88] Barteková, E. and P. Börkey (2022), “Digitalisation for the transition to a resource efficient and circular economy”, OECD Environment Working Papers, No. 192, OECD Publishing, Paris, https://doi.org/10.1787/6f6d18e7-en.
[87] Bauer et al., P. (2021), “A digital twin of Earth for the green transition”, Nature Climate Change, Vol. 11, pp. 80-83.
[25] Benqassem, S. et al. (2021), “Digital technologies in Europe: An environmental life cycle approach”, study commissioned by the European Parliamentary Group of the Greens/EFA, GreenIT.fr, December, https://www.greenit.fr/wp-content/uploads/2021/12/EU-Study-LCA-7-DEC-EN.pdf.
[1] Bieser et al., J. (2023), “A review of assessments of the greenhouse gas footprint and abatement potential of information and communication technology”, University of Zurich, https://www.zora.uzh.ch/id/eprint/228892/1/2023_A%20review%20of?%20assessments%20of%20the%20greenhouse%20gas%20footprint%20and%20abatement%20potential%20of%20information%20and%20c.pdf.
[153] BMK (2021), Austria’s 2030 Mobility Master Plan, Austria Federal Ministry for Climate Action, Environment, Energy, Mobility, Innovation and Technology, Vienna, https://www.bmk.gv.at/en/topics/mobility/mobilitymasterplan2030.html.
[70] Bremer et al., C. (2023), “Assessing energy and climate effects of digitalization: Methodological challenges and key recommendations”, nDEE Framing Paper Series, SSRN, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4459526.
[140] Bundesregierung (2023), “The public sector set to become a role model”, 19 April, Bundesregierung, Berlin, https://www.bundesregierung.de/breg-en/federal-government/the-energy-efficiency-act-2184958.
[92] Casanovas, M. and A. Nghiem (1 August 2023), “Cybersecurity – Is the power system lagging behind?”, IEA Commentaries, https://www.iea.org/commentaries/cybersecurity-is-the-power-system-lagging-behind.
[50] Central Statistics Office (2023), Data Centres Metered Electricity Consumption 2022, 12 June, Central Statistics Office, Cork, https://www.cso.ie/en/releasesandpublications/ep/p-dcmec/datacentresmeteredelectricityconsumption2022/keyfindings.
[146] Cervantes, M. et al. (2023), “Driving low-carbon innovations for climate neutrality”, OECD Science, Technology and Industry Policy Papers, No. 143, OECD Publishing, Paris, https://doi.org/10.1787/8e6ae16b-en.
[38] Chancerel, P. et al. (2015), “Estimating the quantities of critical metals embedded in ICT and consumer equipment”, Resources, Conservation & Recycling, Vol. 98, pp. 9-18, https://www.sciencedirect.com/science/article/abs/pii/S0921344915000488.
[98] ChatClimate (2023), ChatClimate, website, https://www.chatclimate.ai (accessed on 6 February 2024).
[119] Christian-Albrechts-Universität zu Kiel (2023), “Rasmus: Cost-efficient Ship Routing utilizing Ocean Currents”, webpage, https://www.ai.informatik.uni-kiel.de/de/projects/current-projects-2/copy_of_rasmus-ship-routing (accessed on 13 April 2024).
[10] CODES (2022), “Action Plan for a Sustainable Planet in the Digital Age”, United Nations Environment Programme, Nairobi, https://www.unep.org/resources/report/action-plan-sustainable-planet-digital-age.
[59] Crawford, K. (2021), Atlas of AI, Yale University Press, New Haven, CT, https://yalebooks.yale.edu/book/9780300264630/atlas-of-ai.
[120] Creutzig, F. et al. (2015), “Transport: A roadblock to climate change mitigation?”, Science, Vol. 350/6263, pp. 911-912, https://doi.org/10.1126/science.aac8033.
[133] D4S (2022), Digital Reset: Redirecting Technologies for the Deep Sustainability Transformation, Digitalisation for Sustainability, Berlin, https://digitalization-for-sustainability.com/digital-reset.
[49] Danish Energy Agency (2020), Denmark’s Climate and Energy Outlook, Danish Energy Agency, Copenhagen, https://ens.dk/sites/ens.dk/files/Basisfremskrivning/deco_2020_27082020.pdf.
[141] Defra (2020), “Greening government: ICT and digital services strategy 2020-2025”, Policy Paper, UK Department for Environment, Food & Rural Affairs, London, https://www.gov.uk/government/publications/greening-government-ict-and-digital-services-strategy-2020-2025/greening-government-ict-and-digital-services-strategy-2020-2025.
[115] EEA (2022), Transport and Environment Report 2022, European Environment Agency, Copenhagen, https://www.eea.europa.eu/publications/transport-and-environment-report-2022.
[51] EirGrid (2021), All-Island Generation Capacity Statement 2021-2030, EirGrid Plc, Dublin and SONI Ltd, Belfast, https://www.eirgridgroup.com/site-files/library/EirGrid/208281-All-Island-Generation-Capacity-Statement-LR13A.pdf.
[32] EITO (2002), The Impact of ICT on Sustainable Development, European Information Technology Observatory, Madrid.
[118] Elkin, C. and D. Sanekommu (8 August 2023), “How AI is helping airlines mitigate the climate impact of contrails”, Google blog, https://blog.google/technology/ai/ai-airlines-contrails-climate-change.
[56] Ember (2023), Global Electricity Review 2023, Ember, London, https://ember-climate.org/app/uploads/2023/04/Global-Electricity-Review-2023.pdf.
[55] Enerdata (2023), “Electricity domestic consumption”, World Energy & Climate Statistics 2023, Enerdata, https://yearbook.enerdata.net/electricity/electricity-domestic-consumption-data.html.
[81] EUR-Lex (2012), Directive 2012/19/EU of the European Parliament and of the Council of 4 July 2012 on waste electrical and electronic equipment (WEEE) (recast) Text with EEA relevance, Eur-Lex, http://data.europa.eu/eli/dir/2012/19/oj.
[22] European Commission (2023), ICT Task Force Study – Final Report, Publications Office of the European Union, Luxembourg, https://data.europa.eu/doi/10.2760/486253.
[41] European Commission (2023), “Study on the critical raw materials for the EU 2023”, Final Report, Publications Office of the European Commission, Luxembourg, https://op.europa.eu/en/publication-detail/-/publication/57318397-fdd4-11ed-a05c-01aa75ed71a1.
[156] European Commission (2023), “Use of NextGenerationEU Proceeds”, webpage, https://commission.europa.eu/strategy-and-policy/eu-budget/eu-borrower-investor-relations/nextgenerationeu_en#:~:text=NextGenerationEU%20is%20the%20EU’s%20%E2%82%AC,digital%20and%20more%20resilient%20future (accessed on 31 March 2024).
[143] European Parliament (2023), Artificial Intelligence Act: Amendments adopted by the European Parliament on 14 June 2023 on the proposal for a regulation of the European Parliament and of the Council on laying down harmonised rules on artificial intelligence, European Parliament, Strasbourg, https://www.europarl.europa.eu/doceo/document/TA-9-2023-0236_EN.html.
[135] European Union (2021), Regulation (EU) 2021/241 of the European Parliament and of the Council of 12 February 2021 establishing the Recovery and Resilience Facility, Official Journal of the European Union, https://eur-lex.europa.eu/eli/reg/2021/241/oj.
[83] Eurostat (2023), “Waste Statistics – Electrical and Electronic Equipment”, webpage, https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Waste_statistics_-_electrical_and_electronic_equipment#Electrical_and_electronic_equipment_.28EEE.29_put_on_the_market_and_WEEE_processed_in_the_EU (accessed on 31 March 2024).
[155] Federal Chancellery of Switzerland (2023), Digital Switzerland Strategy, Federal Chancellery of Switzerland, Geneva, https://digital.swiss/en/strategy/strategie.html#download.
[138] Finland Ministry of Transport and Communications (2021), Climate and Environmental Strategy for the ICT Sector, Finland Ministry of Transport and Communications, Helsinki, https://julkaisut.valtioneuvosto.fi/bitstream/handle/10024/162912/LVM_2021_06.pdf?sequence=1&isAllowed=y.
[150] Finnish Government (2021), Sustainable Growth Programme for Finland: Recovery and Resilience Plan, Publications of the Finnish Government 2021:69, Helsinki, https://julkaisut.valtioneuvosto.fi/bitstream/handle/10024/163363/VN_2021_69.pdf?sequence=1&isAllowed=y.
[73] Fisher, S. (2023), “Data centers leverage cooling to heat homes”, 18 January, Data Center Knowledge, https://www.datacenterknowledge.com/power-and-cooling/data-centers-leverage-cooling-heat-homes#close-modal.
[53] Fitri, A. (2023), “Inside the data centre moratorium movement”, 29 June, Tech Monitor, https://techmonitor.ai/technology/cloud/inside-the-data-centre-moratorium-movement.
[77] Forti, V. et al. (2020), The Global E-waste Monitor 2020: Quantities, flows and the circular economy potential, United Nations University (UNU)/United Nations Institute for Training and Research (UNITAR) – co-hosted SCYCLE ProgrammeInternational Telecommunication Union (ITU) & International Solid Waste Association (ISWA), Bonn/Geneva/Rotterdam, https://ewastemonitor.info/wp-content/uploads/2020/11/GEM_2020_def_july1_low.pdf.
[47] Freitag, C. et al. (2021), “The real climate and transformative impact of ICT: A critique of estimates, trends, and regulations”, Patterns, Vol. 2/9, p. 100340, https://doi.org/10.1016/j.patter.2021.100340.
[102] G20 (2021), G20 Guidelines for Financing and Fostering High-Quality Broadband Connectivity for a Digital World, Italian G20 Presidency for the G20 Infrastructure Working Group, October 2021, http://www.g20.utoronto.ca/2021/G20-Guidelines-for-Financing-and-Fostering-High-Quality-Broadband-Connectivity-for-a-Digital-World.pdf.
[160] Gartner (2022), “Gartner Glossary: Digital Twin”, webpage, https://www.gartner.com/en/information-technology/glossary/digital-twin (accessed on 19 January 2024).
[42] Geng, Y., J. Sarkis and R. Bleischwitz (2023), “How to build a circular economy for rare-earth elements”, Nature, Vol. 619/7969, pp. 248-251, https://www.nature.com/articles/d41586-023-02153-z.
[137] Germany Federal Ministry for the Environment, Nature Conservation and Nuclear Safety (2020), Digital Policy Agenda for the Environment, Germany Federal Ministry for the Environment, Nature Conservation and Nuclear Safety, Bonn, https://www.bmuv.de/fileadmin/Daten_BMU/Download_PDF/Digitalisierung/digitalagenda_policy_paper_en_bf.pdf.
[54] Google Cloud (2022), “Incident Report”, webpage, https://status.cloud.google.com/incidents/fmEL9i2fArADKawkZAa2 (accessed on 19 January 2024).
[139] Government of France (2021), Feuille de route: Numérique et environnement – faisons converger les transitions, Government of France, Paris, https://www.economie.gouv.fr/files/files/PDF/2021/Feuille_de_route_Numerique_Environnement_vremerciement1802.pdf?v=1663683594.
[136] Government of Korea (2023), “Korean New Deal”, 28 July, Press Release, Ministry of Industry and Finance, Seoul, https://english.moef.go.kr/pc/selectTbPressCenterDtl.do?boardCd=N0001&seq=4948#fn_download.
[151] Government of Mexico (2022), General Law on Climate Change, Government of Mexico, Mexico City, https://www.diputados.gob.mx/LeyesBiblio/pdf/LGCC.pdf.
[74] Green Mountain (2021), “Land-based lobster farming will use waste heat from data center”, 22 June, News Release, Green Mountain, Stavanger, Norway, https://press.greenmountain.no/pressreleases/land-based-lobster-farming-will-use-waste-heat-from-data-center-3107932.
[37] Gupta, U. et al. (2020), “Chasing carbon: The elusive environmental footprint of computing”, arXiv 2011.02839, https://arxiv.org/pdf/2011.02839.pdf.
[26] Hertin, F. and J. Berkhout (2001), “Impacts of information and communication technologies on environmental sustaimability: Speculations and evidence”, report commissioned by the OECD Directorate for Science Technology and Industry, Paris.
[19] Hilty, L. and B. Aebischer (2014), “ICT for sustainability: An emerging research field”, in Advances in Intelligent Systems and Computing, ICT Innovations for Sustainability, Springer International Publishing, Cham, https://doi.org/10.1007/978-3-319-09228-7_1.
[99] Hirn, J. et al. (2022), “A deep generative artificial intelligence system to predict species coexistence patterns”, Methods in Ecology and Evolution, Vol. 13/5, pp. 1052-1061, https://doi.org/10.1111/2041-210X.13827.
[48] IEA (2024), Electricity 2024, IEA, Paris, https://iea.blob.core.windows.net/assets/6b2fd954-2017-408e-bf08-952fdd62118a/Electricity2024-Analysisandforecastto2026.pdf.
[39] IEA (2023), Critical Minerals Market Review 2023, IEA, Paris, https://www.iea.org/reports/critical-minerals-market-review-2023.
[17] IEA (2023), “Data Centres and Data Transmission Networks”, webpage, https://www.iea.org/energy-system/buildings/data-centres-and-data-transmission-networks (accessed on 19 January 2024).
[23] IEA (2023), Energy Technology Perspectives, IEA, Paris, https://www.iea.org/reports/energy-technology-perspectives-2023.
[60] IEA (2023), “Global CO2 emissions by sector, 2019–2022”, Data and Statistics, IEA, Paris, https://www.iea.org/data-and-statistics/charts/global-co2-emissions-by-sector-2019-2022.
[69] IEA (2022), World Energy Outlook 2022, IEA, Paris, https://www.iea.org/reports/world-energy-outlook-2022.
[46] IEA (2021), “Global trends in Internet traffic, data centre workloads and data centre energy use, 2010-2020”, Data and Statistics, IEA, Paris, https://www.iea.org/data-and-statistics/charts/global-trends-in-internet-traffic-data-centre-workloads-and-data-centre-energy-use-2010-2019.
[147] IEA (2021), Net Zero by 2050: A Roadmap for the Global Energy Sector, OECD Publishing, Paris, https://doi.org/10.1787/c8328405-en.
[134] IEEE SA (2023), “Strong sustainability by besign”, webpage, https://sagroups.ieee.org/planetpositive2030/our-work (accessed on 22 February 2024).
[154] INSP (2023), “Tronc commun aux écoles de service public”, webpage, https://insp.gouv.fr/tronc-commun-aux-ecoles-de-service-public (accessed on 19 January 2024).
[28] ISO (2006), “ISO 14040:2006, Environmental management — Life cycle assessment — Principles and framework”, webpage, https://www.iso.org/standard/37456.html (accessed on 19 January 2024).
[29] ISO (2006), “ISO 14044:2006, Environmental management — Life cycle assessment — Requirements and guidelines”, webpage, https://www.iso.org/standard/38498.html (accessed on 19 January 2024).
[116] ITF (2023), ITF Transport Outlook 2023, OECD Publishing, Paris, https://doi.org/10.1787/b6cc9ad5-en.
[14] ITU (2023), World Telecommunication/ICT Indicators Database, International Telecommunications Union, Geneva, https://www.itu.int/en/ITU-D/Statistics/Pages/publications/wtid.aspx.
[114] ITU-T (2023), ITU/WMO/UNESCO IOC Joint Task Force to investigate the use of submarine telecommunications cables for ocean and climate monitoring and disaster warning, webpage, https://www.itu.int/en/ITU-T/climatechange/task-force-sc/Pages/default.aspx (accessed on 30 March 2024).
[27] ITU-T (2018), “L.1450 : Methodologies for the assessment of the environmental impact of the information and communication technology sector”, webpage, https://www.itu.int/rec/T-REC-L.1450-201809-I/en (accessed on 19 January 2024).
[68] Johnson, D. (2022), “Scarcity drives fabs to wastewater recycling: Boosting water recycling at fabs by up to 98 percent keeps chip production on target”, 25 January, IEEE Spectrum blog, https://spectrum.ieee.org/fabs-cut-back-water-use.
[112] Jones, J. (2022), “450MHz network to be built out for critical infrastructure in Germany”, 15 February, Smart Energy International, https://www.smart-energy.com/smart-grid/450mhz-network-to-be-built-out-for-critical-infrastructure-in-germany.
[61] Kaack, L. et al. (2022), “Aligning artificial intelligence with climate change mitigation”, Nature Climate Change, Vol. 12/6, pp. 518-527, https://doi.org/10.1038/s41558-022-01377-7.
[72] Karnama, A., E. Haghighi and R. Vinuesa (2019), “Organic data centers: A sustainable solution for computing facilities”, Results in Engineering, Vol. 4, p. 100063, https://doi.org/10.1016/j.rineng.2019.100063.
[132] Ker, D., P. Montagnier and V. Spiezia (2021), “Measuring telework in the COVID-19 pandemic”, OECD Digital Economy Papers, No. 314, OECD Publishing, Paris, https://doi.org/10.1787/0a76109f-en.
[79] Kumar, A., M. Holuszko and D. Espinosa (2017), “E-waste: an overview on generation, collection, legislation and recycling practices”, Resource Conservation and Recycling, Vol. 122, pp. 32-42, https://doi.org/10.1016/j.resconrec.2017.01.018.
[89] Lange, S. and T. Santarius (2020), Smart Green World? Making Digitalization Work for Sustainability, Routledge, Milton Park, England.
[100] Larosa, F. et al. (2023), “Halting generative AI advancements may slow down progress in climate research”, Nature, Vol. 13, pp. 497-499, https://www.nature.com/articles/s41558-023-01686-5.
[96] Leavitt, M. (18 July 2022), “MosaicML ResNet Part 2: Research Deep Dive”, Databricks blog, https://www.mosaicml.com/blog/mosaic-resnet-deep-dive.
[105] Lenninghan, M. (2021), “KPN links €1 billion credit facility to sustainability goals”, 5 August, Telecoms.com, https://telecoms.com/510827/kpn-links-e1-billion-credit-facility-to-sustainability-goals.
[45] Li, P. et al. (2023), “Making AI less ‘thirsty’: Uncovering and addressing the secret water footprint of AI models”, arXiv 2304.03271, https://doi.org/10.48550/arXiv.2304.03271.
[34] Livingstone, L. et al. (2022), “Synergies and trade-offs in the transition to a resource-efficient and circular economy”, OECD Environment Policy Papers, No. 34, OECD Publishing, Paris, https://doi.org/10.1787/e8bb5c6e-en.
[2] Luo, Y. et al. (2019), “A decision support system for waste heat recovery and energy efficiency improvement in data centres”, Applied Energy, Vol. 250, pp. 1217-1224, https://doi.org/10.1016/j.apenergy.2019.05.029.
[7] Maes, M. et al. (2022), “Monitoring exposure to climate-related hazards: Indicator methodology and key results”, OECD Environment Working Papers, No. 201, OECD Publishing, Paris, https://doi.org/10.1787/da074cb6-en.
[36] Malmodin, J., P. Bergmark and S. Matinfar (2018), “A high-level estimate of the material footprints of the ICT and E&M sector”, ICT for Sustainability, EPiC Series in Computing, Vol. 52, pp. 168-186, https://doi.org/10.29007/q5fw.
[75] McGovan, J. (2023), “Germany sets binding 2030 target for energy consumption reduction”, 22 September, Clean Energy Wire, https://www.cleanenergywire.org/news/germany-sets-binding-2030-target-energy-consumption-reduction.
[128] Michels, L. et al. (2022), “Empowering consumers to make environmentally sustainable online shopping decisions: A digital nudging approach”, Hawaii International Conference on System Sciences Proceedings, 4-7 January, Maui, HI.
[30] Mickoleit, A. (2010), “Greener and Smarter: ICTs, the Environment and Climate Change”, OECD Green Growth Papers, No. 2010/1, OECD Publishing, Paris, https://doi.org/10.1787/5k9h3635kdbt-en.
[149] Ministry of Finance (2021), Denmark’s Recovery and Reslience Plan: Accelerating the Green Transition, Ministry of Finance, Copenhagen, https://en.fm.dk/media/18774/denmarks-recovery-and-resilience-plan-accelerating-the-green-transition_web.pdf.
[11] Muench, S. et al. (2022), “Towards a green and digital future”, Publications Office of the European Union, Luxembourg, https://publications.jrc.ec.europa.eu/repository/handle/JRC129319.
[62] Mytton, D. (2021), “Data centre water consumption”, npj Clean Water, Vol. 4/1, https://doi.org/10.1038/s41545-021-00101-w.
[91] Nafkha-Tayari., W. et al. (2022), “Virtual power plants optimization issue: A comprehensive review on methods, solutions, and prospects”, Energies, Vol. 15/10, p. 3607, https://doi.org/10.3390/en15103607.
[6] NOAA (2024), “2023 was the world’s warmest year on record, by far”, 12 January, National Ocean and Atmospheric Administration, Washington, D.C., https://www.noaa.gov/news/2023-was-worlds-warmest-year-on-record-by-far#:~:text=Earth’s%20average%20land%20and%20ocean,0.15%20of%20a%20degree%20C).
[16] OECD (forthcoming), Review of Relevance of the OECD Recommendation on Information and Communication Technologies and the Environment, OECD Publishing, Paris.
[15] OECD (2024), Broadband Statistics (database), https://www.oecd.org/en/topics/sub-issues/broadband-statistics.html (accessed on 19 January 2024).
[130] OECD (2024), “Household spending”, (indicator), https://doi.org/10.1787/b5f46047-en (accessed on 7 March 2024).
[13] OECD (2024), OECD Digital Economy Outlook 2024 (Volume 1): Embracing the Technology Frontier, OECD Publishing, Paris, https://doi.org/10.1787/a1689dc5-en.
[58] OECD (2023), “A blueprint for building national compute capacity for artificial intelligence”, OECD Digital Economy Papers, No. 350, OECD Publishing, Paris, https://doi.org/10.1787/876367e3-en.
[159] OECD (2023), “Definition of an AI System”, OECD.AI Policy Observatory, https://oecd.ai/en/ai-principles (accessed on 7 March 2024).
[97] OECD (2023), G7 Hiroshima Process on Generative Artificial Intelligence (AI): Towards a G7 Common Understanding on Generative AI, OECD Publishing, Paris, https://doi.org/10.1787/bf3c0c60-en.
[93] OECD (2023), Net Zero+: Climate and Economic Resilience in a Changing World, OECD Publishing, Paris, https://doi.org/10.1787/da477dda-en.
[148] OECD (2023), OECD Science, Technology and Innovation Outlook 2023: Enabling Transitions in Times of Disruption, OECD Publishing, Paris, https://doi.org/10.1787/0b55736e-en.
[57] OECD (2023), “Population”, OECD data explorer (indicator), https://data.oecd.org/pop/population.htm (accessed on 23 April 2024).
[90] OECD (2022), “Broadband networks of the future”, OECD Digital Economy Papers, No. 327, OECD Publishing, Paris, https://doi.org/10.1787/755e2d0c-en.
[5] OECD (2022), “Building better societies through digital policy: Background paper for the CDEP Ministerial meeting”, OECD Digital Economy Papers, No. 338, OECD Publishing, Paris, https://doi.org/10.1787/07c3eb90-en.
[109] OECD (2022), “Communication regulators of the future”, OECD Digital Economy Papers, No. 333, OECD Publishing, Paris, https://doi.org/10.1787/f02209e6-en.
[113] OECD (2022), “Developments in spectrum management for communication services”, OECD Digital Economy Papers, No. 332, OECD Publishing, Paris, https://doi.org/10.1787/175e7ce5-en.
[20] OECD (2022), Measuring the environmental impacts of artificial intelligence compute and applications: The AI footprint, OECD Publishing, Paris, https://doi.org/10.1787/7babf571-en.
[9] OECD (2022), Recommendation of the Council on Blockchain and Other Distributed Ledger Technologies, OECD Legal/0474, Compendium of Legal Instruments, OECD, Paris, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0470.
[101] OECD (2021), Recommendation of the Council on Broadband Connectivity, OECD/Legal 0322, Compendium of Legal Instruments, OECD, Paris, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0322.
[76] OECD (2019), “E-waste generated per capita”, OECD Going Digital Toolkit, https://goingdigital.oecd.org/en/indicator/53 (accessed on 19 January 2024).
[35] OECD (2019), Global Material Resources Outlook to 2060: Economic Drivers and Environmental Consequences, OECD Publishing, Paris, https://doi.org/10.1787/9789264307452-en.
[3] OECD (2019), Going Digital: Shaping Policies, Improving Lives, OECD Publishing, Paris, https://doi.org/10.1787/9789264312012-en.
[12] OECD (2018), “IoT measurement and applications”, OECD Digital Economy Papers, No. 271, OECD Publishing, Paris, https://doi.org/10.1787/35209dbf-en.
[21] OECD (2017), “Information and communication technology (ICT)”, (indicator), https://doi.org/10.1787/04df17c2-en (accessed on 16 March 2024).
[161] OECD (2017), OECD Digital Economy Outlook 2017, OECD Publishing, Paris, https://doi.org/10.1787/9789264276284-en.
[125] OECD (2017), Tackling Environmental Problems with the Help of Behavioural Insights, OECD Publishing, Paris, https://doi.org/10.1787/9789264273887-en.
[144] OECD (2013), “Mobile handset acquisition models”, OECD Digital Economy Papers, No. 224, OECD Publishing, Paris, https://doi.org/10.1787/5k43n203mlbr-en.
[145] OECD (2012), Case Study on Critical Metals in Mobile Devices, 19 September, Working Party on Resource Productivity and Waste, OECD, Paris, https://one.oecd.org/document/ENV/EPOC/WPRPW(2012)4/FINAL/en/pdf.
[4] OECD (2010), Recommendation of the Council on Information and Communication Technologies and the Environment, OECD Legal/0380, Compendium of Legal Instruments, OECD, Paris, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0380.
[18] OECD (2007), “Information Economy – Sector definitions based on the international standard industry classification (ISIC 4)”, OECD, Paris, https://portal.oecd.org/eshare/sti/pc/Pages/Project.aspx?ProjectID=84.
[117] OECD/ITF (2015), Automated and Autonomous Driving, OECD/ITF, Paris, https://www.itf-oecd.org/sites/default/files/docs/15cpb_autonomousdriving.pdf.
[86] OECD-FAO (2021), OECD-FAO Agricultural Outlook 2021-2030, OECD Publishing, Paris, https://doi.org/10.1787/19428846-en.
[82] One Planet Network (2023), Rethinking, Extending, Re-using. Harnessing Digital Technologies for the Circular Economy, One Planet Network Secretariat, UN Environment, Paris, https://www.oneplanetnetwork.org/sites/default/files/2023-11/RETHINKING%2C%20EXTENDING%2C%20REUSING-%20HARNESSING%20DIGITAL%20TECHNOLOGIES%20FOR%20THE%20CIRCULAR%20ECONOMY%20-%205%20%281%29.pdf.
[131] République française (2021), “LOI n° 2021-1485 du 15 novembre 2021 visant à réduire l’empreinte environnementale du numérique en France”, République française, Paris, https://www.legifrance.gouv.fr/jorf/id/JORFTEXT000044327272.
[142] République française (2021), “LOI n° 2021-1755 du 23 décembre 2021 visant à renforcer la régulation environnementale du numérique par l’Autorité de régulation des communications électroniques, des postes et de la distribution de la presse”, République française, Paris, https://www.legifrance.gouv.fr/jorf/id/JORFTEXT000044553569#:~:text=de%20r%C3%A9gulation%20...-,LOI%20n%C2%B0%202021%2D1755%20du%2023%20d%C3%A9cembre%202021%20visant,distribution%20de%20la%20presse%20(1).
[71] ReUseHeat (2017), “Demo Sites: Braunschweig”, webpage, https://www.reuseheat.eu/brunswick (accessed on 24 March 2024).
[84] Rifkin, J. (2022), The Age of Resilience: Reimagining Existence on a Rewilding Planet, Swift Press, London.
[126] Rivers, N. (2018), “Leveraging the Smart Grid: The Effect of Real-Time Information on Consumer Decisions”, OECD Environment Working Papers, No. 127, OECD Publishing, Paris, https://doi.org/10.1787/6ad4d5e3-en.
[85] Rolnick, D. et al. (2022), “Tackling climate change with machine learning”, ACM Computing Surveys, Vol. 55/2, pp. 1-96, https://doi.org/10.1145/3485128.
[33] Science Based Targets (2020), “Information and Communication Technology”, webpage, https://sciencebasedtargets.org/sectors/ict (accessed on 10 January 2024).
[65] Siddik, M., A. Shehabi and L. Marston (2021), “The environmental footprint of data centers in the United States”, Environmental Research Letters, Vol. 16/6, p. 064017, https://doi.org/10.1088/1748-9326/abfba1.
[122] Silva, O. et al. (2022), “Environmental impacts of autonomous vehicles: A review of the scientific literature”, Science of the Total Environment, Vol. 830, https://doi.org/10.1016/j.scitotenv.2022.154615.
[94] Singapore Land Authority (2024), Singapore Geospatial Master Plan, Singapore Land Authority, https://geospatial.sla.gov.sg/initiatives/masterplan.
[123] Sudhakar, S., V. Sze and S. Karaman (2022), “Data centers on wheels: Emissions from computing onboard autonomous vehicles”, IEEE Micro, Vol. 43/1, pp. 29-39, https://doi.org/10.1109/MM.2022.3219803.
[124] Sunstein, C. and L. Reisch (2013), “Automatically green: Behavioral economics and environmental protection”, Harvard Environmental Law Review, Vol. 38/1, https://doi.org/10.2139/ssrn.2245657.
[52] Swindhoe, D. (2022), “UK and European data center colo firms face energy price “turbulence,“ warns FTI”, 27 May, Datacentre Dynamics, https://www.datacenterdynamics.com/en/news/uk-and-european-colo-firm-faces-energy-price-turbulence-warns-fti.
[106] Telecom Review (2020), “KPN plans €3.5 billion fiber rollout by 2024”, 26 November, Telecom Review, https://www.telecomreview.com/index.php/articles/wholesale-and-capacity/4321-kpn-plans-3-5-billion-fiber-rollout-by-2024.
[107] Telefónica (2021), “Telefónica issues the first sustainable hybrid bond of the telco sector, amounting to 1 Billion euros”, webpage, https://www.telefonica.com/en/web/press-office/-/telefonica-issues-the-first-sustainable-hybrid-bond-of-the-telco-sector-amounting-to-1-billion-euros (accessed on 4 November 2021).
[129] Thaler, R. and C. Sunstein (2009), NUDGE: Improving Decisions about Health, Wealth, and Happiness, Penguin, New York.
[66] The New York Times (2021), “Taiwan is facing a drought, and it has prioritized its computer chip business over farmers”, 4 April, The New York Times, https://www.nytimes.com/2021/04/08/business/taiwan-is-facing-a-drought-and-it-has-prioritized-its-computer-chip-business-over-farmers.html.
[24] Thompson, N. et al. (2020), “The computational limits of deep learning”, arXiv 2007.05558, https://doi.org/10.48550/arXiv.2007.05558.
[40] UNCTAD (2020), “Digital economy growth and mineral resources: Implications for developing countries”, UNCTAD Technical Notes on ICT for Development, No. 16, United Nations Conference on Trade and Development, Geneva, https://unctad.org/system/files/official-document/tn_unctad_ict4d16_en.pdf.
[64] United Nations (2022), The United Nations World Water Development Report 2023, UNESCO, Paris, https://www.unwater.org/publications/un-world-water-development-report-2023.
[63] Uptime Institute (2021), Global Data Center Survey 2021, Uptime Institute, New York, https://uptimeinstitute.com/uptime_assets/4d10650a2a92c06a10e2c70e320498710fed2ef3b402aa82fe7946fae3887055-2021-data-center-industry-survey.pdf.
[80] Vishwakarma, S. et al. (2022), “E-waste in information and communication technology sector: Existing scenario, management schemes and initiatives”, Environmental Technology & Innovation, Vol. 27/9, p. 102797, https://doi.org/10.1016/j.eti.2022.102797.
[44] Wäger, P., R. Hischier and R. Widmer (2015), “The material basis of ICT”, in ICT Innovations for Sustainability. Advances in Intelligent Systems and Computing, Springer, Cham.
[31] WBCSD and WRI (2004), GHG Protocol Revised, World Business Council for Sustainable Development, Geneva, and World Resources Institute, Washington, D.C., https://ghgprotocol.org/sites/default/files/standards/ghg-protocol-revised.pdf.
[158] White House (2022), Inflation Reduction Act Guidebook, The White House, Washington, D.C., https://www.whitehouse.gov/cleanenergy/inflation-reduction-act-guidebook.
[108] WIK-Consult (2020), “Neutral fibre and the European Green Deal”, report commissioned by Stokab, WIK-Consult, October, https://stokab.se/download/18.15d457b6178eff38ee02ed/1619701526100/Neutral%20fibre%20and%20the%20European%20Green%20Deal%20,%20WIK-Consult.pdf.
[43] Willenbacher, M. and V. Wohlgemuth (2022), “Rebound effects in the use of rare earth metals in ICT”, International Journal of Environmental Sciences, Vol. 30/1, p. 556277, https://doi.org/10.19080/IJESNR.2022.30.556277.
[8] WMO (2023), WMO Global Annual to Decadal Climate Update – Target Years: 2022 and 2022-2026, World Metereological Organization, Geneva, https://library.wmo.int/idurl/4/56318.
← 1. Please refer to OECD ongoing work on the “Review of Relevance of the OECD Recommendation on Information and Communication Technologies and the Environment” (OECD, forthcoming[16]).
← 2. The OECD has an ongoing cross-directorate project on the twin “digital” and “green” transitions.
← 3. Indicators based on the Universal Thermal Climate Index (UTCI). The dataset provides a global assessment of changes in average temperature by providing the temperature anomaly, days with above-average temperatures and days with below-average temperatures. Extreme temperature indicators have been prepared jointly by the OECD and IEA, see the reference paper for a more complete description of the methods.
← 4. “The ICT sector combines manufacturing and services industries whose products primarily fulfil or enable the function of information processing and communication by electronic means, including transmission and display. The ICT sector contributes to technological progress, output and productivity growth. Its impact can be examined in several ways: directly, through its contribution to output, employment or productivity growth, or indirectly, as a source of technological change affecting other parts of the economy, for instance.” (OECD, 2017[161]).
← 5. A digital twin is “a digital representation of a real-world entity or system. The implementation of a digital twin is an encapsulated software object or model that mirrors a unique physical object, process, organisation, person or other abstraction. Data from multiple digital twins can be aggregated for a composite view across a number of real-world entities, such as a power plant or a city, and their related processes” (Gartner, 2022[160]).
← 6. For the purpose of this chapter, “ICTs” and “digital technologies” are used interchangeably.
← 7. See OECD (2007[18]) for the full categorisation of ICT products and services. While ICTs have been a subject of measurement and study at the OECD for several decades, the terms “digital technologies” and “ICTs” are often used interchangeably as a precise definition of both terms has not been universally adopted. That said, the OECD’s definition of ICTs has been instrumental in current measurement standards on the impact of ICTs and the environment, such as the efforts by the International Telecommunication Union Telecommunication Standardization Sector (ITU-T, 2018[27]). The lack of harmonised definitions can limit evidence-based analysis, especially when comparing statistics on the environmental impacts of digital technologies.
← 8. According to the OECD definition of an AI system, updated in late 2023, an AI system is “a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment” (OECD, 2023[159]).
← 9. The inclusion of certain device categories can lead to the upper bound estimates. For example, end-user devices such as smartphones and TVs make up a significant part of the ICT GHG footprint depending on methodologies and definitions used.
← 10. The authors refer to the continent Americas, including all countries in North, Central and South America.
← 11. This report put forward a number of measures that decision makers should consider for achieving two main goals: i) to increase collection of mobile devices, instead of generating waste, and ii) to develop environmentally sound management (ESM) systems for waste in developing countries with large informal sectors (OECD, 2013[144]; OECD, 2012[145]).
← 12. An exchange rate of EUR 0.950/USD for the year 2022 from OECD.stat has been used.