Data centers 2021: Data center boom in Germany continues - Cloud computing drives the growth of the data center industry and its energy consumption

Abstract

The energy consumption of data centers continues to increase. At 17 billion kWh, data centers consumed 6.5 % more electricity in 2021 than in 2020. The main reason for the growth in energy consumption is the expansion of cloud data centers in Germany and the associated increase in the number of large data centers. However, traditional data centers operated by companies themselves also continue to have a high share of data center capacities in Germany.

Full text

Data centers 2021
Data center boom in Germany continues
Cloud computing drives the growth of the data
center industry and its energy consumption
Dr Ralph Hintemann | Simon Hinterholzer
The energy consumption of data centers continues to in-
crease. At 17 billion kWh, data centers consumed 6.5 %
more electricity in 2021 than in 2020. The main reason
for the growth in energy consumption is the expansion
of cloud data centers in Germany and the associated in-
crease in the number of large data centers. However,
traditional data centers operated by companies them-
selves also continue to have a high share of data center
capacities in Germany.
These are the results of a recent study by the Borderstep
Institute on the development of the energy consumption
of data centers in Germany.
The growth in the energy consumption of data centers is
mainly due to the increased demand for IT performance
from data centers. Although servers are becoming more
and more powerful and the utilization of IT systems is in-
creasing, especially in cloud data centers, the sharp in-
crease in power demand is leading to higher electricity
consumption in data centers in Germany (figure 1). Of the
17 billion kWh of electricity consumed in data centers in
2021, almost 11 billion kWh went to IT components (serv-
ers, storage and network). Since 2010, the electricity de-
mand of the IT components in the data centers has thus
almost doubled. In comparison, the energy consumption
of the data center infrastructure (cooling/uninterruptible
power supply (UPS), etc.) increased only moderately from
4.6 to 6.1 billion kWh between 2010 and 2021. Therefore,
the efficiency of the data center infrastructure has im-
proved significantly. The so-called PUE value1 is used as a
measure of the efficiency of the data center infrastruc-
ture. The average PUE value of data centers and small IT
installations in Germany improved from 1.82 to 1.56 be-
tween 2010 and 2021.
1 The Power Usage Effectiveness (PUE) value indicates the ratio of
the annual energy demand of the entire computing centre to the
annual energy demand of the IT of the computing centre.
Figure 1: Energy consumption of servers and data centers in
Germany from 2010 to 2021 (Source: Borderstep)
If current trends continue, the energy consumption of
data centers in Germany will continue to rise in the future
despite the significant efficiency gains in IT and infrastruc-
ture components. By 2030, an increase to about 28 billion
kWh is expected. (Hintemann, Graß, Hinterholzer &
Grothey, 2022).
Servers need more and more power
A major reason for the rising energy consumption is the
increasing power consumption of the servers. If we look
at the 2-CPU servers measured for the SPECPower®
benchmark, they have a maximum power consumption
today that is almost two and a half times as high as in 2007
(figure 2). Despite this increase in power consumption, to-
day's systems are many times more efficient when com-
pared to the increase in computing power. The maximum
computing power of the systems has increased by a factor
of 50 in some cases over the same period.

Figure 2: Maximum power consumption of servers with two
CPUs in the SPECpower® benchmark (2007 - 2021)
Strong growth of cloud data centers
According to current research by Eurostat, the share of
companies (10 employees or more) in Germany using
cloud services will be 42% in 2021, at the average Euro-
pean level. Especially in the Scandinavian countries Swe-
den (75 %), Finland (75 %), Denmark (65 %) and Norway
(64 %), cloud computing is used significantly more. But
also, in the Netherlands (65%) and Italy (60%) there is a
significantly higher share of companies using cloud ser-
vices. However, cloud use in Germany has been increasing
significantly in recent years. Between 2018 and 2021
alone, the number of cloud-using companies almost dou-
bled (Eurostat, 2022). One reason for this is not least the
Corona pandemic. Cloud providers in particular benefited
from the digitalization boost associated with the pan-
demic in the past year and will probably also emerge
stronger from the crisis in the long term. (eco & Arthur D.
Little, 2020). The significant growth in demand for cloud
services is one reason for the boom in the construction of
large data centers in Germany, which has been going on
for years. The cloud currently accounts for one-third of
data center capacity in Germany. By 2025, cloud compu-
ting is expected to become the dominant deployment
model in Germany. Especially in the Frankfurt Rhine-Main
and Berlin/Brandenburg regions, there are currently a
number of new data center projects with rated IT power
in the order of 100 MW and more. (Hintemann et al.,
2022).
Own operation of data centers continues to be very
important
Despite the strongly increasing share of cloud computing
in data center capacities in Germany, the in-house opera-
tion of data centers is also an important issue for a large
proportion of companies. The capacities of traditional
data centers in Germany have so far remained at a con-
2 An article on this topic for a scientific journal is currently being pre-
pared.
3 When mining cryptocurrencies, high computing power is usually
used in a network to ensure that the currencies are tamper-proof.
stant level. The growth in total capacities is almost exclu-
sively due to cloud data centers. It is also to be expected
that this development will change little in the future. Ac-
cording to a recent representative survey, only about 5 %
of companies are planning to do without their own data
centers. (Hintemann et al., 2022).
Waste heat from data centers must be recovered for
climate protection
Against the backdrop of advancing climate change and
the challenge of reducing dependence on fossil fuels such
as oil and gas, the use of waste heat is becoming increas-
ingly important. Data centers can make a significant con-
tribution to this. Up to now, the electricity consumed in
data centers has been released unused into the environ-
ment in the form of heat. Until now, this was due in par-
ticular to the lack of economic efficiency of waste heat
utilization from data centers. (Clausen, Hintemann &
Hinterholzer, 2020). In the future, however, it can be as-
sumed that the conditions will be significantly different,
so that the use of waste heat from data centers can be-
come very important. The reasons for this are the chang-
ing framework conditions of energy supply, new technical
approaches such as liquid cooling of servers or the use of
artificial intelligence to optimize heat extraction and heat
networks.
Worldwide development: High share of crypto-min-
ing in the increasing energy demand of data centers
If one analyses the available studies and publications on
the worldwide energy consumption of data centers, no
uniform picture emerges. Some researchers assume a sig-
nificant increase in energy consumption worldwide (An-
drae, 2019; Belkhir & Elmeligi, 2018; Petit, Carlini &
Avelar, 2021; The Shift Project, 2019). In other studies, on
the other hand, the energy consumption of data centers
has remained almost constant in recent years. (IEA, 2017;
Masanet, Shehabi, Lei, Smith & Koomey, 2020; Shehabi,
Smith, Masanet & Koomey, 2018). For the year 2020, for
example, the calculations range from about 200 billion
kWh to almost 900 billion kWh.
With the help of the structural bottom-up model of data
centers developed for Germany, an estimate of the devel-
opment of the global energy consumption of data centers
can be made. This can be done on the basis of worldwide
sales figures for servers and by adjusting the model as-
sumptions. Taking into account uncertainties due to pa-
rameter variations, a global energy consumption of data
centers in the order of 350 to 500 billion kWh/a can be
calculated in 20212 . Mining3 of cryptocurrencies is re-
sponsible for about a quarter of this energy consumption.
If one analyses various model calculations on the subject
During this process, new "coins" are created ("mining") and distrib-
uted to the miners as a reward for their computing power.

of crypto-mining (CBECI, 2022; Digiconomist, 2022a,
2022b) an energy consumption in the order of 80 to 120
billion kWh/a seems likely.
Methodology of the study
This study is based on work by the Borderstep Institute on
the development of data centers in Germany.
According to the underlying system, data centers are all
closed spatial units such as server cabinets, server rooms,
parts of buildings or entire buildings in which IT compo-
nents such as servers, storage and network components
are installed. Small IT installations that centrally provide
computing and storage services are also explicitly consid-
ered data centers. The development of data center capac-
ities is calculated in particular on the basis of the server
equipment in the data centers. The different performance
classes of servers are also taken into account.
The calculations are made with the help of a comprehen-
sive structural bottom-up model of the data center land-
scape in Germany and Europe, which was developed at
the Borderstep Institute and is updated annually (Fichter
& Hintemann, 2014; Hintemann, 2017a, 2020; Hinte-
mann, Clausen, Beucker & Hinterholzer, 2021; Hinte-
mann, Fichter & Stobbe, 2010; Hintemann et al., 2022;
Hintemann & Hinterholzer, 2019, 2020; Hintemann, Hin-
terholzer, Montevecchi & Stickler, 2020; Hintemann et al.,
2020; Stobbe et al., 2015). The model describes the data
centers in various size classes in terms of their equipment
with different server types, storage systems and network
infrastructures. The age structure of the servers and the
energy requirements of the various server types in differ-
ent operating states are also taken into account. In addi-
tion, the data center infrastructures such as air condition-
ing, power supply, UPS, etc. are modelled.
The following sources in particular were used for the cur-
rent calculations:
 Study " Rechenzentren in Deutschland - Aktuelle
Marktentwicklungen " (Hintemann et al., 2022)
 Energy-efficient Cloud Computing Technologies and
Policies for an Eco-friendly Cloud Market" study
(Hintemann et al., 2020)
 Study "Data Centers in Europe - Opportunities for
Sustainable Digitalization - Part 1" (Hintemann &
Hinterholzer, 2020)
 Study " Entwicklung des IKT-bedingten Strombedarfs
in Deutschland " - Study by Fraunhofer IZM and Bor-
derstep on behalf of the Federal Ministry for Eco-
nomic Affairs and Energy (Stobbe et al., 2015)
 Current results of studies on the development of the
data center market (CBRE, 2017, 2018, 2020; Cisco,
2015, 2016; Gartner, 2020; Hintemann, 2014, 2017b;
Hintemann & Clausen, 2018a, 2018b; Hintemann,
Fichter & Schlitt, 2014; Howard-Healy, 2018)
 Data from the market research institute Techconsult
on the market development of server, storage and
network components (eanalyzer) (Techconsult, 2014,
2015, 2016)
 Data from the market research institutes IDC and
EITO on the market development for servers in Ger-
many and Europe (EITO, 2014, 2019; IDC, 2018, 2021)
 Scientific literature and manufacturer information on
the development of energy consumption of services,
storage and network products and other efficiency
technologies for data centers.
Sources:
Andrae, A. S. G. (2019). Projecting the chiaroscuro of the electricity use
of communication and computing from 2018 to 2030.
Belkhir, L. & Elmeligi, A. (2018). Assessing ICT global emissions foot-
print: Trends to 2040 & recommendations. Journal of
Cleaner Production, 177, 448-463.
CBECI. (2022). Cambridge Bitcoin Electricity Consumption Index
(CBECI). Accessed 16.6.2022. Available at:
https://ccaf.io/cbeci/index
CBRE. (2017). European Data Centres Market Review. Q4 2016. Lon-
don. Accessed 10.6.2021. Available at:
https://www.cbre.de/de-de/research/European-Data-Cen-
tres-MarketView-Q4-2016
CBRE. (2018, March 21). European data centre capacity in demand
again exceeds 100 MW. European data centre capacity in
demand exceeds 100 MW again. Accessed 4.4.2018. Availa-
ble at: http://news.cbre.de/nachgefragte-leistung-eu-
ropaischer-rechenzentren-ubersteigt-erneut-100-mw-wert
CBRE. (2020). Europe Data Centres Q4 2019. Accessed 19.3.2020. Avail-
able at: https://www.cbre.de/en/global/research-and-re-
ports/featured-reports-global/featured-reports-emea
Cisco. (2015). Cisco Global Cloud Index: Forecast and Methodology
2014-2019. Available at: http://www.cisco.com/c/en/us/so-
lutions/collateral/service-provider/global-cloud-index-
gci/Cloud_Index_White_Paper.pdf
Cisco. (2016). Cisco Global Cloud Index: Forecast and Methodology
2015-2020. Accessed 10.10.2017. Available at:
https://www.cisco.com/c/dam/en/us/solutions/collat-
eral/service-provider/global-cloud-index-gci/white-paper-
c11-738085.pdf
Clausen, J., Hintemann, R. & Hinterholzer, S. (2020). Wirtschaftlichkeit
der Abwärmenutzung aus Rechenzentren in Deutschland.
Background paper. Berlin: Borderstep Institut.
Digiconomist. (2022a). Bitcoin Energy Consumption Index. Accessed
16.6.2022. Available at: https://digiconomist.net/bitcoin-en-
ergy-consumption
Digiconomist. (2022b). Ethereum Energy Consumption Index. Accessed
23.5.2022. Available at: https://digiconomist.net/ethereum-
energy-consumption
eco & Arthur D. Little. (2020). The Internet Economy in Germany 2020-
2025: Impact of the Corona Crisis. Accessed 10.5.2020.
Available at: https://www.eco.de/studie-internetwirtschaft-
20-25-corona-preprint/
EITO. (2014). EITO Costumized Report for Borderstep. Berlin: EITO.
EITO. (2019). EITO Customized Report for Borderstep. Berlin: EITO.
Eurostat. (2022). Cloud computing services. Accessed 29.7.2022. Avail-
able at: https://ec.europa.eu/eurostat/data-
browser/view/isoc_cicce_use/default/bar?lang=en
Fichter, K. & Hintemann, R. (2014). Beyond Energy: Material Stocks in
Data Centers, Taking Resource Efficiency into Account in
Green IT Strategies for Data Centers. Journal of Industrial
Ecology, (forthcoming). https://doi.org/DOI:
10.1111/jiec.12155

Gartner. (2020, March 19). Gartner Says Worldwide Server Revenue
Grew 5.1% in the Fourth Quarter of 2019, While Shipments
Increased 11.7%. Gartner. Accessed 19 Mar. 2020. Available
at: https://www.gartner.com/en/newsroom/press-re-
leases/2020-03-19-gartner-says-worldwide-server-revenue-
grew-5-percent-in-the-fourth-quarter-of-2019-while-ship-
ments-increased-11-percent
Hintemann, R. (2014). Consolidation, Colocation, Virtualization, and
Cloud Computing - The Impact of the Changing Structure of
Data Centers on Total Electricity Demand. In L. Hilty & B.
Aebischer (Eds.), ICT Innovations for Sustainability. Advan-
ces in Intelligent Systems and Computing. Berlin, Heidel-
berg: Springer.
Hintemann, R. (2017a). Rechenzentren 2016. Trotz verbesserter Ener-
gieeffizienz steigt der Energiebedarf der deutschen Rechen-
zentren im Jahr 2016. Berlin: Borderstep Institute for Inno-
vation and Sustainability. Available at: https://www.border-
step.de/wp-content/uploads/2017/03/Borderstep_Re-
chenzentren_2016_Stand_07_03_2017_finaln-1.pdf
Hintemann, R. (2017b). Energieeffizienz und Rechenzentren in Deutsch-
land – weltweit führend oder längst abgehängt? - Presenta-
tion. Berlin: Netzwerk energieeffiziente Rechenzentren -
NeRZ. Accessed 25.10.2017. Available at: https://www.bor-
derstep.de/wp-content/uploads/2017/07/NeRZ-Studie-Re-
chenzentrumsmarkt-30-06-2017.pdf
Hintemann, R. (2020). Data centres 2018. Efficiency gains are not
enough: Energy demand of data centres continues to rise
significantly. Berlin: Borderstep Institute. Available at:
https://www.borderstep.de/wp-content/up-
loads/2020/03/Borderstep-Rechenzentren-2018-
20200324rev.pdf
Hintemann, R. & Clausen, J. (2018a). Bedeutung digitaler Infrastruktu-
ren in Deutschland. Sozioökonomische Chancen und Heraus-
forderungen für Rechenzentren im internationalen Wettbe-
werb. Berlin. Available at. Berlin. Accessed 14.6.2018. Avail-
able at: https://www.eco.de/wp-content/uploads/dlm_up-
loads/2018/06/DI_Studie.pdf
Hintemann, R. & Clausen, J. (2018b). Potenzial von Energieeffizienz-
technologien bei Colocation Rechenzentren in Hessen. Ber-
lin: Borderstep Institute for Innovation and Sustainability.
Accessed 26.4.2018. Available at: https://www.digitalstrate-
gie-hessen.de/rechenzentren
Hintemann, R., Clausen, J., Beucker, S. & Hinterholzer, S. (2021). Studie
zu Nachhaltigkeitspotenzialen in und durch Digitalisierung in
Hessen . Study for Hessen Trade & Invest GmbH on behalf of
the Hessian State Chancellery, Hessian Minister for Digital
Strategy and Development. Wiesbaden: Hessian State Chan-
cellery, Minister for Digital Strategy and Development.
Hintemann, R., Fichter, K. & Schlitt, D. (2014). Adaptive computing and
server virtualization in German data centres - Potentials for
increasing energy efficiency today and in 2020. In Marx
Gómez, Sonnenschein, Vogel, Winter, Rapp, & Giesen (Eds.),
Proceedings of the 28th Conference on Environmental Infor-
matics - Informatics for Environmental Protection, Sustaina-
ble Development and Risk Management (pp. 477-484). Ol-
denburg: BIS. Accessed 25.1.2015. Available at: http://envi-
roinfo.eu/sites/default/files/pdfs/vol8514/0477.pdf
Hintemann, R., Fichter, K. & Stobbe, L. (2010). Materialbestand der Re-
chenzentren in Deutschland-Eine Bestandsaufnahme zur Er-
mittlung von Ressourcen- und Energieeinsatz. Study within
the framework of the UFO Plan project "Product-related ap-
proaches in information and communication technology"
(funding code 370 893 302), commissioned by the Federal
Environment Agency.
Hintemann, R., Graß, M., Hinterholzer, S. & Grothey, T. (2022). Rechen-
zentren in Deutschland - Aktuelle Marktentwicklungen.
Bitkom. Accessed on 15.4.2022. Available at:
https://www.bitkom.org/Bitkom/Publikationen/Re-
chenzentren-in-Deutschland-2022
Hintemann, R. & Hinterholzer, S. (2019). Energy Consumption of Data
Centers Worldwide - How will the Internet become Green?
Lappeenranta, Finland. Accessed 8 Aug. 2019. Available at:
http://ceur-ws.org/Vol-2382/ICT4S2019_paper_16.pdf
Hintemann, R. & Hinterholzer, S. (2020). Data Centres in Europe - Op-
portunities for Sustainable Digitalisation - Part 1. Berlin: Alli-
ance to Strengthen Digital Infrastructures in Germany. Ac-
cessed on 11.6.2020. Available at: https://digitale-infra-
strukturen.net/studie-nachhaltige-digitalisierung-in-europa/
Hintemann, R., Hinterholzer, S., Montevecchi, F. & Stickler, T. (2020).
Energy-efficient Cloud Computing Technologies and Policies
for an Eco-friendly Cloud Market. Berlin, Vienna: Borderstep
Institute & Environment Agency Austria. Available at:
https://op.europa.eu/en/publication-detail/-/publica-
tion/bf276684-32bd-11eb-b27b-01aa75ed71a1/language-
en/format-PDF/source-183168542
Howard-Healy, M. (2018). Co-location Market Quarterly (CMQ) brief -
presentation at BroadGroup's Knowledge Brunch in Frank-
furt. Broadgroup.
IDC. (2018). Server Market and Enterprise Storage Systems By Country
2014-2017.
IDC. (2021). IDC Quarterly Enterprise and Cloud Infrastructure Trackers
- Custom Market Intelligence - Prepared for Borderstep.
IEA. (2017). Digitalization & Energy. Accessed 20 Jan. 2019. Available
at: https://www.iea.org/reports/digitalisation-and-energy
Masanet, E., Shehabi, A., Lei, N., Smith, S. & Koomey, J. (2020, February
28). Recalibrating global data centre energy-use estimates.
Science. Accessed 4.3.2020. Available at: https://science.sci-
encemag.org/content/367/6481/984
Petit, V., Carlini, S. & Avelar, V. (2021). Digital economy and climate im-
pact: A bottom-up forecast of the IT sector energy consump-
tion and carbon footprint to 2030. Nanterre: Schneider Elec-
tric. Available at: https://www.se.com/ww/en/in-
sights/tl/electricity-4-0/digital-economy-and-climate-impact
Shehabi, A., Smith, S. J., Masanet, E. & Koomey, J. G. (2018). Data cen-
tre growth in the United States: decoupling the demand for
services from electricity use. Environmental Research Let-
ters, 13(12).
Stobbe, L., Hintemann, R., Proske, M., Clausen, J., Zedel, H. & Beucker,
S. (2015). Entwicklung des IKT-bedingten Strombedarfs in
Deutschland - Study commissioned by the Federal Ministry
for Economic Affairs and Energy. Berlin: Fraunhofer IZM and
Borderstep Institute. Available at:
http://www.bmwi.de/BMWi/Redaktion/PDF/E/entwick-
lung-des-ikt-bedingten-strombedarfs-in-deutschland-ab-
schlussbericht,property=pdf,bereich=bmwi2012,spra-
che=de,rwb=true.pdf
Techconsult. (2014). Data from the eanalyzer. Available at: www.eana-
lyzer.biz
Techconsult. (2015). Data from the eanalyzer. Available at: www.eana-
lyzer.biz
Techconsult. (2016). Data from the eanalyzer. Available at: www.eana-
lyzer.biz
The Shift Project. (2019). LEAN ICT- Towards digital sobriety. Accessed
18 Apr. 2019. Available at: https://theshiftpro-
ject.org/en/article/lean-ict-our-new-report/
Contact:
Dr Ralph Hintemann
Partner and Senior Researcher
Borderstep Institute for Innovation and Sustainability
Non-profit limited liability company
Clayallee 323
14169 Berlin, Germany
Tel. +49 (0)30 306 45-1005
E-mail: hintemann@borderstep.de
www.borderstep.de