Internet ’ s handprint

In this decade there will an unprecedented growth of generated data, computations, instructions, and operations. This growth may not compromise clean air, clean water and a sustainable energy and material usage, but rather facilitate these prerequisites for flora and fauna. There are many indications (expected trends and estimates) showing that the Internet Sector will be able to provide solutions to other Sectors such as Buildings, Transportation and Industry which will help reduce the total global consumption of energy and materials. For instance, products are replaced by virtual services e.g. by using e-readers instead of paperbacks, and transportation is avoided by online shopping or Internet meetings. This is more resource and energy efficient than before and entire sectors, like transport, industry, and agriculture can be optimized. Internet may foster new sustainable lifestyles which can lower the affluence despite certain rebound effects. The underlying idea is that e.g. human-related global greenhouse gas (GHG) supply can be significantly halted if existing and developing ICT Solutions are used in other sectors (and in the Internet infrastructure itself) to cause a handprint. Such solutions include products-sold-as-services, smart Grid and smart metering. Compared to earlier approaches, the 2020 transformative effects on smart work, land use and smart circularity are included in the discussion, as well as consequential LCA modelling. Internet’s handprint will be 4-7 times its footprint in 2030. The handprint is highly dependent e.g. on how large share of the buildings can adopt smart metering and the product to service rate. Internet will in itself use intelligent ICT solutions as well as neuromorphic, reversible and superconducting computing as well as nanophotonics to mitigate its own material and energy use. However, more importantly the intelligent ICT solutions should be used in the rest of the society to reach efficiency goals. Power saving is a highly efficient strategy for cost reduction in the Internet Sector itself and beyond.


Introduction
I n the present decade there will be an unprecedented growth of generated data, computations, instructions and operations. This growth may not compromise clean air, clean water and a sustainable energy and material usage, but rather facilitate these prerequisites for flora and fauna. Overall, global primary energy consumption rises due to the Internet [1] and Internet's own electrical energy consumption is also rising [2][3][4]. Plausibly the global primary energy and electricity consumption would rise even faster without the handprint of certain ICT solutions. The Internet Sector is one of few which might off-set its own electrical energy consumption and GHG supply, i.e., its handprint [5] is larger than its footprint. There are many indications (expected trends and estimates) showing that the Internet infrastructure will be able to provide solutionswithin main Sectors such as Buildings, Transportation and Industry -which will help reduce (halt the increase of) the total global consumption of energy and materials [6]. Information and communications technologies (ICTs) can potentially contribute to reduce resource consumption through increased productivity in many Sectors by enabling total optimization and dematerialization, occasionally using artificial intelligence (AI) [7]. Internet's deflationary characteristics suggest that it has a handprint. AI and machine learning (ML) are cornerstones of intelligent ICT Solutions which make them unique compared to incremental improvements. For instance, products are replaced by virtual services, e.g., by using e-readers instead of paperbacks [8], 81 transportation is avoided by online shopping or online chatting. This is more resource and energy efficient than before and entire sectors, like transport, industry, and agriculture can be optimized. Internet may foster new sustainable lifestyles which can lower the affluence despite certain rebound effects. E.g. e-reader adopters are yet to fully abandon paper books for e-books suggesting a total net increase [8]. The underlying idea is that e.g. total anthropogenic global GHG supply (TAGGHGS) can be significantly halted if existing and developing ICT Solutions are used in other sectors (and Internet itself [9]) to their "full potential" in a smart manner. Such solutions include Products-sold-as-Services, Smart Grid and Smart Metering. The avoidance potential is highly dependent e.g. on how large share of the building GHG supply can be reduced by Smart Metering [10,11].
Internet will use smart ICT solutions to keep its own material and electrical energy use under control. The smart ICT solutions could also be used in the rest of the society to reach environmental goals. Still, the Internet Sector itself has a huge responsibility to try to reach high annual electrical energy efficiency gains of ≈ 20% in data centers and networks. This seems to have been the case in the last decade. New technologies such as neuromorphic, reversible and superconducting computing as well as nanophotonics may help in this decade [12,13]. Several attributional life cycle assessments have shown that Internet's share of TAGGHGS may have been stable 2015-2020 [13]. However, although recent literature is divided [14], the trends of rampant instructions/second and slowing improvements of switching energy are very clear [12,14]. Related cryptocurrency mining electrical energy demand is on the rise but not necessarily the related GHG supply [15]. Therefore power saving is a highly efficient strategy for GHG supply reduction in the Internet itself and beyond. In this work, the potential TAGGHGS avoidance of using ICT Solutions for energy saving, compared to low adoption of ICT Solutions, is explored. An algorithm for the estimation is established.

Objectives
The objective of this prediction study is to estimate the changes (GHG is proxy) to occur between 2019 and 2030 if traditional technologies are replaced with Information and Communication Technology (ICT) technologies. Internet's scope according to [16,17] consists of the use stage of end-user consumer devices, network infrastructure and data centers as well as the production of hardware for all. The attributional LCA approach [16,17] may not be able to capture the actual GHG avoidance derived from the use of ICT solutions, as many of them have the ability to decrease the energy and material losses. Consequently, less energy and materials need to be produced and purchased by a final customer in order to consume the same quantity of product. A consequential LCA (CLCA) with a planetary system boundary is attempted for ICT solutions handprint.

Hypotheses
The hypothesis is that Internet's GHG supply will increase according to the expected scenario as outlined by Andrae [16]. Moreover, Internet's GHG supply is off-set already in 2020 by ICT Solutions and the handprint will be 6 times the footprint by 2030.

Materials and methods
The approach for estimating Internet's direct GHG supply is established while the handprint potential of ICT solutions for TAGGHGS is less clear. Here, for the sake of modelling, the World is divided into seven sectors -Industry, Buildings, Transport, Travel, Agriculture, Waste and Land use. Then several ICT Solutions ability to reduce TAGGHGS in each sector is estimated. The approach is very much simplistic as there are highly granular Input-Output models [9,18] which describe the economic flows of different sectors in the society. Therefore, the coupling of IO and LCA can be applied to model indirect impacts of changes in product inputs and outputs in several economic sectors [19]. The coupling of IO and LCA can cover all economic sectors in a large geographical boundary. All assumptions made are available in the Supplementary Information.

Description of method for estimating Internet GHG supply
The approach for Internet direct GHG supply follows the one outlined in [16] expected scenario. Table 1 shows some global trends assumptions derived from [13,16]. Trends for TAGGHGS are followed closely [20].

Handprint -description of method for estimating GHG supply reductions and power savings by ICT Solutions
The overall methodological approach for estimating electricity demand and GHG supply handprint by the Internet in other Sub-Sectors (Industry, Transportation, Buildings, Agriculture, Land use and Waste) is described below. Land use leads to increased GHG supply if new plants (e.g. trees) are not planted which can absorb CO 2 . Waste (management) is a relevant Sub-Sector of its own, e.g. landfill, recycling, incineration etc.

Consequential handprint LCA
The functional unit of the CLCA is: global demand of electricity and TAGGHGS. Figure 1 shows the principle of provided functions replacing more inefficient ways. ICT Solutions are replacing traditional solutions in the CLCA. Increased digitalization leads to increased production of ICT Solutions which substitute products, energy, materials and land use. When less travel and transport are used, also less fuel is produced. To compensate for "missing" fuel, more electricity will be produced. These effects contribute to the view of Internet driving deflation.

Estimation of total anthropogenic global GHG supply
The GHG supply from the Sub-Sector Agriculture in 2030 is ≈ 8.4 Gt as shown in Table 2 below. Table 2 shows that Internet's share of TAGGHGS without Internet handprint will be low (≈ 2%). Moreover, without active Internet handprint, Industry's GHG supply will increase almost 20% between 2020 and 2030. Land use and waste environmental impacts are more challenging to reduce with ICT Solutions, but they still contribute to TAGGHGS which could rise between 2020 and 2030 [13].
Despite economic growth and rebound effects, TAGGHGS could likely be significantly halted if existing and future ICT Solutions are used in other sectors (and Internet itself). Table 2 shows the present estimations by Sector for approximate TAGGHGS.
Equation (2) below shows the total global handprint of ICT Solution j in year t: where, TAGGHGS t = Total anthropogenic GHG supply in year t; GHG i,t = Anthropogenic GHG supply from Sector type i in year t; ICT hp,t = ICT Solutions total global GHG handprint in year t; j = ICT Solution type; i = Sector type; t = year; MVCV j,i,t = Marginal Variable Change Vector of ICT Solution j in Sector i in year t; F j,i,t = Fraction of Sector i which is applicable to ICT Solution j in year t.
Here follows two examples which explain somewhat (2); • In 2030, F is 0.1 for Travel Sector for "Video/telemeeting, air" as it is assumed that 10% of the Travel sector GHG supply are air travel GHG supply which can be reduced (50%, MVCV = 0.5) via "Video/telemeeting, air" ICT solutions. The minimum values for this case are F = 0.01 and MVCV = 0.05. • In 2030, F is 0.5 for Building Sector for "Smart metering in Buildings" as it is assumed that 50% of the Building sector GHG supply are electricity related GHG supply which can be reduced (10%, MVCV = 0.1) via "Smart metering in Buildings" ICT solutions. The minimum values for this case are F = 0.05 and MVCV = 0.01. MVCV Smart metering, Buildings,2020=0.03 is reported in literature [21]. Table 2 shows that the GHG supply from Industry is ≈ 18 Gt in 2020 with around 6 Gt related to electricity demand. Transport Electricity GHG and Travel Electricity GHG are both ≈ 6% of respective Sectors total GHG, while Land use Electricity GHG is excluded as it is assumed close to zero. However, with the anticipated electrification of vehicles, both Transport and Travel Electricity GHG will increase [17].

Division of GHG supply between electricity and other sources for Industry, Buildings, Travel and Transports
Moreover, hydrogen production for fuel cell vehicles -and indirectly ammonia production for internal combustion engines -will add to the electricity demand of the Transport and Travel Sectors [13]. Hydrogen production will also add electricity demand in the Industry sector (e.g. Steel supply chain), but at the same time the net GHG supply may be reduced in e.g. Steel production [13].
Ammonia has potential for Travel and Transport as a more or less non-ICT based GHG Supply reducer as ammonia can be used in converted internal combustion engines [22]. It is especially useful to know Industry Electricity Use and Building Electricity Use to understand the effectiveness of different electrical energy efficiency strategies such as Smart Metering. Internet also has some GHG supply from other sources than electricity, such as from diesel generators [23] but they have been excluded. Table 3 shows the split between electricity GHG and other sources for each Sector. Table 3 may help understand where electrification, hydrogen and ammonia solutions make the most sense in energy related activities.

Estimation of ICT Solutions handprints
They key question addressed here is: How much GHG supply can smart ICT Solutions help avoid, ICT hp,t , in other sectors of society each year between t = 2019 and t = 2030? Despite large uncertainties, quite likely more GHG supply can be avoided than the Internet emits itself i.e., ICT hp,t > GHG i=Internet,t , perhaps already for t = 2022. Table 4 shows the addressed Fraction (F) of each Sector (i) and how much the ICT technology (j) can reduce (MVCV). Table 5 roughly outlines how ICT hp,t could increase year by year from 2019 to 2030 as estimated in the present study to reach (at the most) ≈ 11Gt in 2030. However, likely some reduction of TAGGHGS has already occurred historically due to ICT solutions and sensitivity checks are performed in Section 4. Table 5 shows mainly future potential to 2030. In Sections 2.5.1 to 2.5.9 the numbers in Table 5 are explained. Table 6 outlines roughly how electricity handprints (TWh) could increase linearly year by year from 2020 to 2030 as estimated in the present study to reach ≈ 8497 TWh in 2030. As soon as 2022 more TWh can be cut by the Internet than its own usage. Table 7 shows three examples of where Smart Grid can achieve transformation. Smart Grid savings are firstly that 50% of all Buildings GHG supply are applicable for 10% reduction each using Smart Metering. This means that 0.1 × 0.5 × 16.2 Gt= 0.81 Gt can be avoided in 2030. Smart Metering makes the users aware of the power consumption [21]. Smart Grid savings can also be obtained from "Power grid optimization" which may reduce 50% (power load balancing) of power use in each case it is introduced, but perhaps only 10% of global electricity GHG supply is applicable for optimization in 2030, i.e., 0.

Smart Grid
Further Smart Grid related savings by fewer losses than traditional grids [26] are possible from facilitation of renewable energy sources. This is assumed to reduce 10% of the power used in each applicable case which may be 20% of all globally used electricity, 0.1 × 0.2 × 34718TWh × 0.000533Gt/TWh = 0.37Gt. AI-driven battery management (including self-repair) is part of the renewable energy story for ICT handprint. All in all in 2030, 0.81 + 0.92 + 0.37 = 2.1Gt savings in 2030 from Smart Grid. Table 8 shows an example in which Smart Agriculture can achieve transformation. Smart Agriculture can be used to reduce TAGGHGS by e.g. more surveillance and less manual inspection. It is estimated that 10% of GHG supply of the entire Agriculture sector can be reduced, 0.10 × 8.43Gt = 0.84Gt. AI is especially useful in Agriculture [28]. Autonomous variable herbicide spraying can save > 50% of liquid applied per hectare [29].

Smart Services
Smart Services is a rather wide concept for Smart ICT but mainly it is about virtualization and dematerialization. The link to servitization is very strong [29,30]. Table 9 shows two examples in which Smart Services can achieve transformation. E-readers and audio books are examples of book-as-a-service (potentially) replacing physical books [8]. Global Change Mix Factors [32] may reveal to which degree this will materialize and then cause a rebound effect on F. Regarding financial products, digital solutions on blockchain may change the electricity demand of the financial systems [33]. Another example of smart service is e-commerce [34]. With the current rate of digitalization it seems likely that 100% of all Industry products could be sold as services in 2030. It is assumed that 10% of all travel transport and 10% of all Industry GHG supply can be avoided by selling products as services. All in all, 0.1 × 1 × 4.6Gt + 0.1 × 1 × 20.7Gt = 2.5Gt.

Smart Travel
Smart Travel is about optimizing travel routes and vehicle sharing. It is assumed that three main mechanisms lead to savings; smart public travel, fleet car management and route optimization. Table 10 shows two examples in which Smart Travel solutions can achieve transformation.  [37] 50% [37] AI Taxi is an example of fleet car management [7]. The eighth and ninth handprints may be overestimated. Map route travel service handprints are already achieved. All in all, Smart Travel may reduce GHG supply by (0.

Smart Work
Smart Work is about reducing business travel and commuting and the need for less hotel rooms and offices. In 2020 the airline passenger traffic shrunk 67% compared to 2019 and was reduced to 1999 levels [38]. No matter the reason, this suggests that digitalization tools could reach a very high implementation (high F) for video/telemeetings already in 2020 (Table 4). Table 11 shows five examples in which Smart Work solutions can achieve transformation. The tenth and eleventh services could be similar and double counted. A sensitivity analysis will include such issues. All in all, Smart Work may reduce GHG supply by (0.5 × 0.1 + 0.5 × 0.1 + 0.5 × 0.2) × 4.6Gt + (0.2 × 0.1 + 0.1 × 0.01) × 16.2Gt = 1.26Gt.

Smart Buildings
Smart ICT is facilitating automated heating, ventilation and air conditioning (HVAC) systems as well as light control. Via deep learning and cloud-based computing, ICT solutions autonomously optimizes existing HVAC control systems for lowest possible energy consumption. Table 12 shows two examples in which Smart Building solutions can achieve transformation.  [40] 10% [45] Office energy use here, and Smart metering in Section 2.5.1, may address somewhat similar flows. HVAC savings are however also related to thermal energy and not only electricity. All in all, Smart Building solutions may reduce GHG supply by (0.5 × 0.3 + 0.2 × 0.1) × 16.2Gt = 2.75Gt.

Smart transports
Smart Transports is mainly about optimization of truck logistics and shifting transport from e.g. car to train [46] . Table 13 shows two examples in which Smart Transport solutions can achieve transformation.  [48] 10% [48] All in all Smart Transport solutions may reduce GHG supply by (0.2 × 0.5 + 0.5 × 0.1) × 4.5Gt = 0.68Gt. Arguably F for ICT solution (Table 13, last row) may be lower than 10%. Another example of a Smart Transport solution is wireless vehicle-vehicle communication with Cooperative Adaptive Cruise Control which saves truck fuel [49].

Smart Circularity
Around 99% of everything that we buy becomes waste after 6 months. Some 2 billion tonnes of waste (garbage) is generated annually of which ≈ 2.5% is e-waste. Total Material Consumption per capita is also increasing. The effective material flow is much higher than the conventional weight flow. The so called Total Material Requirements per kg metal is increasing, i.e., the ore grades are diminishing. Ore grades (e.g. copper) is gradually decreasing 2.5% per annum, while production and energy consumption (and GHG supply) from mining is increasing [50]. Using AI software for optimization is likely a more fruitful route than new waste management technologies. In product design, AI may predict product design variables for GHG supply reduction and customer relevance. AI can help the Waste sector by Smart logistics (improvements in route planning), mobile collection of e-waste on demand [51], and via intelligent optical sorting machines.
Here the example of intelligent Optical sorting machines is used to exemplify how much the GHG supply from the Waste Management Sector can be reduced. AI helps capturing data from optical sorters from which machinery can learn and "make" decisions that optimize sorting [52]. AI-Powered Robot Picking is another example of Intelligent Circularity. Smart Circularity can be used in Waste Management to avoid GHG supply. Waste can be identified for its proper handling. It is assumed that in each case Smart Circularity is used 10% of the GHG supply can be avoided and that all waste related GHG supply can be addressed by intelligent sorting in 2030, i.e., 0.1 × 1 × 2.1Gt = 0.21Gt. One could argue that selling services instead of products is also smart circularity [31]. Nevertheless, Industry related material savings (servitization) is addressed by Smart Services in Section 2.5.3.
Route planning of garbage trucks is addressed by Smart Transport in Section 2.5.7. Table 14 shows an example in which a Smart waste management solution can achieve transformation. i Transport j MVCV F AI enabled optical sorting 10% [52] 100% [52] All in all Smart Circularity solutions may reduce GHG supply by 0.1 × 1 × 2.1 Gt = 0.21Gt.

Smart land use
AI can be used for managing sustainable land use [53]. Smart Forestry (IoT monitoring) can be used to reduce GHG supply caused by illegal tree cutting [54,55]. Table 15 shows an example in which a Smart land use solution can achieve transformation.

Total Internet GHG supply handprint
All in all, Internet's handprint will be 11.37 Gt in 2030 using assumed (very high) values for MVCV and F. However, when introducing smaller minimum values for MVCV and F, the handprint will be much smaller, but still safely higher than Internet's footprint.

Estimation of global GHG supply with and without ICT handprint and Share of electric power GHG of total global GHG
It is relevant to estimate to which degree the savings from ICT Solutions of Table 5 can reduce TAGGHGS  (Table 2). In Table 16 is shown how TAGGHGS is slowed down. In 2020 the TAGGHGS is 0.2% (57.4 instead of 57.5) less thanks to ICT. In 2030 TAGGHGS could be around 17% (54.9 instead of 66.3) lower than business as usual. Table 16 shows the estimation of global GHG supply with and without ICT handprint and share of electric power GHG supply.

Handprint in each Sector
In this section the savings made possible by Smart Grid, Smart Travel, Smart Buildings etc. are allocated to each societal sector (Table 17). For example the savings by Smart Work are allocated to Travel if the Smart Work savings are travel related, and savings by Smart Grid are allocated to Buildings if the Smart Grid savings are Buildings related.

Share of each sectors GHG supply that can be cut by ICT Solutions
This section discusses the share of each sectors' GHG supply which could be cut year by year from 2019 to 2030. The shares are obtained by dividing the Handprint in the Sector (Table 17) with the Sector (Table 2). Table 18 the shares of each sectors GHG supply that can be cut over time are shown.

Effect of Internet handprint on global GHG supply
In Table 19 the handprint of ICT solutions on TAGGHGS is summarized. In 2030 at the most ≈ 17% of TAGGHGS (11.37 Gt of 66.3 Gt) can be reduced. In Section 4 it is elaborated with sensitivity analysis what speaks for 11.37 Gt and what speaks against.

Specific ICT Solution share of total handprint in 2030
Here specific ICT Solutions are listed in Table 20 and their share of total handprint. The largest reduction amount, 2.43 Gt and 21%, can be achieved from "Office energy use" and "Selling products as services, servitization" 2.07 Gt and 18%.

Individual products handprint
In the present decade there will be an unprecedented growth of generated data shown in Table 21 [16]. This can be used to estimate Internet's handprint per Byte. Individual product's services handprints can then be estimated if their bandwidths are known. Evidently, applicable handprint services need to have specific pathways for creating handprints.

Results
The order of magnitude of the GHG Supply handprint is reasonable. Figure 2 is a graphical representation of the present article. It seems plausible that Internet's handprint will be higher than its footprint. However, 11.37 Gt may be heavily overestimated.

Discussion
What would overthrow the results in the present research? Are there any new developments since 2015 which would falsify magnificent Internet handprints? What is only theory and what are the facts? Is here Internet's handprint massively exaggerated several times? One obvious issue is to which extent the digitalization has already been achieved, meaning that F's max value for t=2030 is already achieved. 2020 meant a huge leap forward in this respect for some ICT Solutions such as "Video/telemeetings, air" and "Video/telemeetings, car". Anyway the reduction of GHG supply observed in 2020 is not really caused by certain ICT solutions. However, they made e.g. working from home possible. Here the view is taken that ICT solutions number 10 and 11 in Table 4 are tested big-scale making the assumed 50% (F=0.5) cut of applicable Transport and Travel GHG Supply between 2019 and 2020 possible.
On the other hand it is not argued that Internet's handprint will lead to absolute smaller TAGGHGS, but rather a slow-down of the increase of TAGGHGS. 2020 will show this fact e.g. with fewer travels and more virtual meetings. 2020 showed that digital ICT technologies can be widely adopted very quickly. Smart Buildings is a very important area for ICT solution offsetting of TAGGHGS as Buildings use much energy. However, it may be argued that "Smart Metering in Buildings" and "Smart Buildings office energy use reduction" is equal. Also, old buildings may not be able to reduce energy use as much as newly constructed ones. These limitations are considered to be included in the sensitivity analysis in which F and MVCV are reduced 10 times. MVCV Smart metering, Buildings,2020 = 0.03 [20] measurement is a good indication for the reasonableness of MVCV Smart metering,Buildings,2030 = 0.1. Likewise, certain Smart Travel ICT Solutions may be similar to Smart Services for Travel. Regarding Smart Work ICT Solutions, "telemeeting" and "teleworking" seem identical with regards to Travel handprint.
If the biggest one -teleworking -is churned the total ICT potential is reduced by 0.48 Gt in 2030. Anyway, for Smart Work the anticipated 2030 savings may already have happened in 2020 due to the global macro changes in 2020. Temporary or permanent decline to a new baseline of airline and automotive travel and transport are the most obvious observations in 2020.
The effectiveness of AI to achieve savings depends greatly on sufficient data and the data scientists and engineers developing the AI software. AI has successfully been employed for forecasting the volume of waste which will be generated. This facilitate proper planning of landfill sites, recycling units, development as well as operation of garbage collection infrastructure. AI can cope especially well with historical data which are of nonlinear nature. Still several indications of savings exist.

ICT Product handprints
What is the link to ICT product related handprints? Dividing the total handprint in Table 5 (11.37 Gt) with the Global Data Center IP Traffic in Table 21 (268 ZettaByte) gives e.g. 0.039 kg CO2e/GigaByte for 2030. However, this intensity is quite rough but may be tested together with specific GigaByte/s bandwidth data.

The right performance for the right application
What performance is good enough for a certain application? Such questions are valid e.g. for Travel Electricity Use (Table 3) where Na-ion batteries (90-115 Wh/kg) could be enough for certain electric vehicles instead of Li-ion (100-265 Wh/kg) [57,58].
The global material efficiency/waste problem seems not to be solved effectively by improved local waste management (e.g. collection). Perhaps an AI optimization of total global supply chains -which targets waste minimization in production and Total Material Consumption/capita is more effective. Optimizing and predicting the whole nonlinear global societal system with Internet as a driver -markets, Input-Output, GHG supply, resources, costs, jobs, waste -is a daunting task which theoretically could better be managed with AI and humans instead of humans alone.

Sensitivity analyses
Without sensitivity checks by 2030 the handprint/footprint ratio will be around 7 (11.37/1.71). Using Monte Carlo simulation and maximum and minimum values of F, MVCV etc. gives an uncertainty spread such that GHGi=Internet,t=2020 = 1.17 Gt (Min 1.07 Gt Max 1.29 Gt) and ICThp,2020 = 0.35 Gt (Min 0.23 Gt Max 0.48 Gt). All assumptions are found in the Supplementary Information. Videomeetings represent almost 50% of Internet's handprint in 2020 but just 4% in 2030 where instead Office energy use and Servitization of products dominate.
Using maximum and minimum values of F, MVCV etc. (see Supplementary Information) gives an uncertainty spread such that GHG i=Internet,t=2030 = 1.53 Gt (Min 1.28 Gt Max 1.8 Gt) and ICT hp,2030 = 6.32 Gt (Min 4.41 Gt Max 8.52 Gt). Compared to the original mean value (11.37 Gt), including the spread of input values reduces the mean value of ICT h p, 2030 substantially due to ten times lower minimum values of F and MVCV. In 2020 the GHG supply by the Internet is some 70% higher than Internet's handprint. Hence, the GHG supply by the Internet is not off-set already in 2020 by ICT Solutions. Nevertheless, the handprint/footprint ratio will be around 4 (6.32/1.53) by 2030. Thus the handprint hypotheses in Section 1.2 are falsified.

Next steps
The present research confirms that the largest handprints from the Internet in 2030 will be in Buildings and Industry. Still, individual nations and service providers would like to estimate better specific handprints. Moreover, the measurable entities in the system should be identified more carefully for individual ICT Solutions such those related to 5G [59]. Additionally, the standardization of the handprint calculation for electronic products should be attempted.
Conflicts of Interest: "The author declares no conflict of interest."