Introduction

Digitalization is permeating all areas of life, with new media consumption patterns emerging, and domestic appliances being expected to be continuously online and accessible. However, digitalization also has negative environmental impacts due to its resource-intensive nature: Production and usage of digital products and services are energy intensive and mostly driven by non-renewable raw materials such as fossil fuels, and the increased production of electronic devices leads to increased amounts of e‑waste (WBGU 2019). The IT sector’s energy demand is projected to increase (Andrae and Edler 2015), and data on its resource demand is scarce (Lutter et al. 2022). To address these issues, the German Environment Agency initiated the DigitalRessourcen project in 2020. The project analyzed the resource demand of digitalization in Germany at micro- and macro-levels through case studies and input-output models and provided insights for resource-efficient digitalization.

Resource consumption in digitalization can be categorized into direct and indirect effects (Hilty and Aebischer 2015). Direct effects refer to the life cycle of ICT devices, while indirect effects consider changes in human behavior and structural changes induced by digitalization. Studies on direct effects often reveal that the total resource demand of a product exceeds the resources physically bound in the product itself (Hilty and Aebischer 2015; Köhler et al. 2018). However, assessing the total resource consumption is typically not feasible when performing macro-level estimations of sector-specific resource consumption since data sets typically do not offer comprehensive consistent material flow information on a global scale (Hintemann et al. 2010; Malmodin et al. 2018). In addition, research has mainly focused on ICT devices for personal use and is only starting to analyze the resource consumption of data centers, communication networks, and emerging technologies (Gröger et al. 2021). Therefore, further research is needed to assess the aggregated resource consumption of digitalization across sectors and provide quantitative information.

In the project DigitalRessourcen, both micro-level assessments based on life cycle analyses with consideration of the resource intensity of individually selected goods and services, as well as macroeconomic analyses using a global input-output database, were conducted. These analyses were carried out independently of each other to generate the most comprehensive empirical basis for assessing the resource intensities of digitalization in Germany. Based on the assessments, fields of actions with levers and further research needed for improving the resource efficiency of digitalization were identified. This article presents the approach and core results of the project.

Assessing resource consumption of digitalization on the micro-level

To assess the resource consumption of digitalization at the micro-level, ten case studies of individual digital services considered highly relevant for digitalization in Germany were examined. The assessment was generally guided by life cycle assessment (LCA) principles according to DIN EN ISO 14040/44 (Deutsches Institut für Normung e. V. 2006). However, the method was refocused by highlighting the inventory level as well as corresponding inputs to focus on analyzing raw material intensities, especially digitalization-relevant materials. The emphasis of the analysis was placed on the manufacturing and use phase. The end-of-life phase was only included in selected case studies and in a simplified manner.

Considering feasibility and data availability, specific indicators were used to calculate direct effects of the investigated use cases of digitalization (see Fig. 1).

Fig. 1: Set of indicators selected for the analysis in the case studies. Source: authors’ own compilation

RMI and TMR were chosen as indicators because they are widely accepted in the scientific field and the calculation method developed by Mostert and Bringezu (2019) and Pauliuk (2022) is readily available for OpenLCA and EcoInvent, the software and database used for the LCA calculations. CED was chosen as an abundantly analyzed indicator for resource use that is highly correlated to GWP and thus strengthened the focus. WDP and LOP were chosen to add a dimension of resource intensity besides raw materials indicators. For the calculation of WDP, LOP and GWP the ReCiPe (H) 2016-method was used, whilst CED was calculated using the method implemented in OpenLCA 1.11 (GreenDelta n.d.). Besides calculating and analyzing the indicators, a focus was set on quantifying the use of bulk raw materials and digitalization-relevant raw materials – a set of materials that are essential for digitalization, as well as low in abundance, global reserves, and availability.

Here, the outcomes of the case study video conferencing are described in more detail as an example. Participating in a one-hour videoconference was found to have a primary RMI of around 116 g, mainly attributable to equipment manufacturing. Gangue rocks (50.5 g), coal (17.9 g), gravel (7.2 g), shale (6.4 g), and sand (2.3 g) were among the significant material inputs. Other materials consumed include calcite (2.0 g), crude oil (1.7 g), clay (1.4 g) and iron (1.2 g). An overview of the most pertinent digitalization-relevant raw materials allocated to the end-user devices is provided in Fig. 2a. The other indicators are also dominated by the manufacturing phase (see Fig. 2b). At the same time, video conferencing offers the potential to reduce environmental impact by avoiding emissions from personal mobility. Studies suggest that remote working and digital solutions can reduce transport and energy emissions by around 15% in 2030 (Lutter et al. 2017).

Fig. 2: a Proportions (%) of digitalization-relevant raw materials in the videoconferencing system (mg per hour of videoconferencing) in the manufacturing phase. b Comparison of the shares for manufacturing and use phase of the calculated indicators for a one-hour video conference. For modeling the usage of the devices at home the German electricity mix from 2018 was used and for modeling the operation of the videoconferencing servers, the global electricity mix from 2018 and 2019 was used to represent a wide set of video conferencing providers (both according to Ecoinvent v3.8). Source: authors’ own compilation

Overarching results and discussion of the case studies

The case studies show diverse drivers of resource use and greenhouse gas potential. Depending on the use case, most resource requirements arise either in the manufacturing phase (e.g., video conferencing, and 3D printing) or in the use phase (e.g., smart home, cryptocurrency, and e‑sports). Manufacturing phase resource requirements are driven by materials and associated processes, with metal ores as dominant inputs. Important raw materials included coal, gravel, shale, sand, crude oil, and specific materials like gallium, tantalum, gold, silver, tin, nickel, lithium, and scandium. The use phase is characterized by electricity demand, driven by the use of fossil fuels such as lignite and hard coal in generation. Gangue is the most abundant material in the use phases. Overall, the results show a correlation between raw material demands, energy input, and global warming potential. These findings align with the literature reviewed for the case studies.

Assessing wider impacts and trends from digitalization

Status quo for 2020 and historic trends

In the macroeconomic analyses, all uses of ICT goods and services (intermediate production inputs, domestic final demand and export purposes) were considered as direct effects of digitalization. The statistical differentiation of ICT goods and services from other economic products was carried out in accordance with OECD’s Guide to Measuring the Information Society (OECD 2011). Based on this guideline, the NACE classification was used to categorize direct digitalization-relevant economic activities in the macroeconomic analyses. Specifically, the following NACE 2 Division and Groups were selected: “Manufacture of computer, electronic and optical products” (26.1, 26.2, 26.3, 26.4, 26.8), “Wholesale of information and communication equipment” (46.5), “Software publishing” (58.2), “Telecommunications” (61), “Computer programming, consultancy and related activities” (62), “Data processing, hosting and related activities; web portals” (63.1), “Repair of computers and communication equipment” (95.1). For the macroeconomic analyses, not all relevant values could be taken directly from the GLORIA database and extensive data work had to be carried out in preparation for the macroeconomic analyses to allocate NACE groups to corresponding larger aggregates.

To quantify the environmental impacts of economy-wide digitalization trends, the Raw Material Input of digitalization (RMI_Dig.) and the Raw Material Consumption of digitalization (RMC_Dig.) in Germany were calculated. RMI reports for a given economy on all raw materials extracted domestically plus all direct raw material imports as well as any raw materials that have already been used along the respective supply chains involved in the production of imported goods and services. RMC reports on that subset of the RMI that is actually used domestically, i.e. raw materials that are neither directly nor indirectly used for the production of exported goods or services. To the authors’ knowledge, RMI_Dig. and RMC_Dig. were calculated for the first time in DigitalRessourcen.

In the macroeconomic analyses, all uses of ICT goods and services […] were considered as direct effects of digitalization.

Overall, RMI_Dig. increased by almost 13.5% between 2000 and 2020 (Fig. 3). With a total value of more than 157 million tons in 2020, it exceeds RMC_Dig. in Germany by more than 60 million tons.

Fig. 3: RMI of the use of digitalization-related goods and services in Germany. Product groups are categorized according to the GLORIA database (Lenzen et al. 2022). Source: authors’ own compilation

The RMI_Dig. is predominantly driven by the use of ICT hardware (Fig. 3): While contributions of ICT services (all groups except ICT hardware) to the German RMI_Dig. declined slightly between the years 2000 and 2020, the use of ICT hardware (areas colored in red) increased this indicator by more than 20 million tons over the same period. In 2020, the use of ICT hardware accounted for more than 81% of the total global raw material use caused by the digitalization of the German national economy.

Future trends under different scenarios

The identified macroeconomic development trends provided inputs for modelling future effects of digitalization in Germany up to the year 2050. For scenario analyses, a novel structural assessment model (GRAMOD) was developed and applied.

Assuming an increase in global population figures and an approximate doubling of global GDP until 2050, the current trends scenario projects a doubling of global production levels for ICT products between 2020 and 2050. For Germany’s RMI_Dig., this scenario projects a slight long-term increase towards a level of close to 160 million tons in 2050. Based on this trend projection, the extent to which the development of the analyzed footprint indicators would be affected by alternative assumptions was simulated regarding

• future efficiency improvements in the domestic production of ICT hardware,

• private households’ consumption propensity,

• the distribution of consumer expenditures on individual goods and services.

Comparing individual simulation results reveals that Germany’s RMI_Dig. could follow different trajectories until 2050: Stronger demand dynamics from private household consumption results in the RMI_Dig. exceeding its 2050 reference value from the trend scenario by more than 12 million tons in the More, Bigger, Faster scenario. Weaker demand dynamics of private households assumed in the Less, Softer, and Greener scenario (with assumed additional efficiency improvements that are rather negligible for the overall development) reduces the RMI_Dig. by almost 6 million tons in 2050 compared to the trend scenario (Fig. 4). All remaining scenario developments did not fall outside the range of these extreme scenarios. For more information on the scenarios, we refer to the forthcoming final public project report.

Fig. 4: Alternative scenario assessments of future developments of the RMI indicator of the use of digitalization-related goods and services in Germany. Source: authors’ own compilation

The results (Fig. 4) highlight that a long-term reduction of Germany’s RMI_Dig. appears to be challenging. This is attributable to the strong international integration of the German economy: All other things being equal, robust global economic growth fuels global demand for German exports, which in turn increases German intermediate demand for digitalization-related services to produce the exported products.

Levers for resource-efficient digitalization

These fields of action relate to the entire life cycle of ICT goods and services, i.e., production, demand, and use, as well as recycling. They can also be correlated to the principles of sustainability – efficiency, consistency, sufficiency.

More specifically, the project results indicate a need for focused initiatives to promote efficiency and reduce environmental impact. Research should focus on designing efficient ICT goods, identifying alternative raw materials, optimizing recycling processes, and improving energy efficiency in data centers and logistics. However, limited and inconsistent data on resource demand hinders measurement. Standardized metrics and systematic data collection are needed (Nilashi et al. 2023). Investigating rebound effects and behavior changes is crucial. A circular-economy approach can contribute to sustainable resource management (Smol et al. 2020). Longitudinal studies, sustainable assessment frameworks, and investment in eco-efficient technologies are necessary. Policymakers should promote renewable energy, circular-economy principles, and digital inclusion. A collaborative effort is required for sustainable digitalization.

Overall, the identified fields of action and related measures aimed to provide initial insights for shaping a sustainable digitalization. Specific measures need to be further detailed, analyzed, and evaluated.

Conclusion

This research article briefly introduced the methods and exemplary results of the DigitalRessourcen project assessing the resource demand of digitalization in Germany. The project revealed the diverging drivers of global resource extractions that are triggered by the digitalization in Germany. Besides domestic demand dependencies, the strong connection of Germany’s economy with international trade could be a main hinderance in reducing future resource demands.

Besides domestic demand dependencies, the strong connection of Germany’s economy with international trade could be a main hinderance in reducing future resource demands.

In addition to the described results, the project also revealed gaps in available data for micro-level assessments, amplified by a lack of harmonized approaches to data collection and indicator calculation. This requires a comprehensive inventory and assessment of available data with targeted identification of deficiencies. This should include investigations on macro-level into which sectors will have an increased demand for digitalization-relevant raw materials in the future and how, for example, a circular economy system can buffer the high demand for digitalization-relevant raw materials.

Lastly, shaping digitalization in a sustainable way and identifying and implementing effective measures is a challenge that needs to be considered in a global context and addressed by a multi-stakeholder dialogue, including politics, industry, and research.