Decarbonizing

Data
Strategic considerations for data center
sustainability in the AI era.

A SEAGATE TECHNOLOGY REPORT

WITH RESEARCH BY DYNATA AND FIELDWORK BY CURRENT GLOBAL APRIL 2025

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 1 |
 Table of Contents
Introduction ................................................................................................................... 03
About the Report ......................................................................................................... 04
Key Takeaways ............................................................................................................... 05

Section 1: Challenges to Decarbonizing Data Centers ............................... 09

Section 2: Solutions That Reduce Data Centers’
Carbon Footprint ................................................................................... 15
Section 3: Rethinking Data Center Sustainability .......................................... 18
Section 4: Turning Challenges Into Opportunities ....................................... 21

Conclusion: The Path to a Sustainable

Datasphere ............................................................................................... 27

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 2 |
 Introduction
As widespread adoption of data- centers, Seagate has a front-row seat
intensive artificial intelligence (AI) to the conversations around data
technologies speeds up worldwide, center sustainability and growing
data center operations are facing demand for cost-efficient data storage.
unprecedented energy demands and Among other things, working with our
a growing carbon footprint. They are customers has taught us that TCO
grappling with escalating environmental and sustainability goals are not
concerns while also striving to mutually exclusive—often they are one
scale compute power and storage and the same.
capacity and to meet their total cost of To better understand the link between
ownership (TCO) goals. TCO and sustainability, we decided
According to McKinsey1, the power to survey data center professionals.
needs of data centers are expected to This report, informed by their insights,
triple by the end of 2030. reveals a critical gap in understanding
“Skyrocketing compute and data data centers’ sustainability challenges,
demands are being further accelerated and the ways these challenges affect
by gains in computing capabilities the entire supply chain. These insights
alongside reductions in chip efficiency can inform decisions that support
relative to power consumption,” the both business growth and
analyst firm noted. environmental goals.
As the global data storage leader It need not be either or.
supplying the world’s leading data

1 Data centers and AI: How data centers and the energy sector can sate AI’s hunger for power, McKinsey, 2024.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 3 |
About the Report
This Seagate Technology report is based on a global quantitative survey.
a commissioned global study conducted The survey gathered responses from 330
by independent research firm Dynata, data center professionals across 11 markets:
with fieldwork by global communications Australia, China, France, Germany, India,
consultancy Current Global, part of The Japan, North America, Singapore, South
Weber Shandwick Collective. Korea, Taiwan, and the United Kingdom.
The study included both qualitative and All respondents work for companies that
quantitative research to examine the AI- manage at least 50 terabytes of storage,
intensified strain on data center sustainability with most overseeing up to five petabytes.
and opportunities for increasing the Participants included CIOs, CTOs, IT VPs,
efficiency of operations. directors, executives, COOs, line-of-business
leaders, storage architects, and solution
The research began with in-depth qualitative
architects in data storage companies and
interviews with five senior experts in data
organizations that supply or provide data
storage and infrastructure from the United
storage solutions.
States, Germany, China, and Japan. Each
expert has over a decade of experience The study explored the current landscape
in data center planning, operations, and of data center efficiency and sustainability. It
sustainability, which enabled robust insights aims to provide industry leaders with data-
into industry challenges and emerging driven insights to guide business-boosting,
trends. These insights informed the design of sustainable decision-making.

Seagate Technology is a global leader in scalable mass-capacity data storage, delivering

more than four and a half billion terabytes of capacity over the past 45 years.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 4 |
Key Takeaways
Data centers are the backbone of potentially accounting for 8% of global sustainability goals. The survey on which
today’s AI economy. Modern-day rapid carbon emissions, up from 0.3% in 20222.   this report is based found that:
adoption of AI technologies has resulted
To keep pace with AI innovation, • Environmental impact is a concern
in data centers facing unprecedented
organizations must scale their data for nearly 95% of respondents.
energy demands and a growing carbon
infrastructure while aligning with • But only 3.3% of the surveyed
footprint. The survey’s results make two
corporate sustainability mandates. data center professionals said
things abundantly clear:
Increasingly, businesses face their companies prioritize low
• 94.5% of the survey respondents expectations—and sometimes regulatory environmental impact in data center
reported that their companies face requirements—to operate data centers purchasing decisions.  
increasing data storage needs. both cost-efficiently and sustainably. It
Still, other findings from this survey
• 97% anticipate AI’s growth to further can be challenging to reconcile these
clarify that TCO and sustainability are far
impact storage demand. two expectations because the more
from competing priorities. The TCO and
energy data centers use, the trickier it is
This corresponds to a growing energy sustainability considerations can and
to reduce their carbon footprints.  
demand. According to data from IEEE: do often align, offering opportunities
by 2030, data center energy demand For this reason, total cost of ownership for operational efficiencies and
is projected to grow significantly, (TCO) is sometimes seen as at odds with environmental impact reduction.

The study found that 1. Energy consumption. High energy infrastructure (identified by 28.5%
usage drives both operational costs survey takers) and acquisition costs
data center operations and carbon emissions, with 53.5% for data center components (27%
are strongly affected of respondents identifying this as a of respondents) significantly impact
by five factors that significant concern.   capital expenditure (CapEx).
embody both TCO and 2. Raw material requirements. 5. Life cycle extension. Over 92% of
sustainability goals. Nearly 49.5% of the surveyed data respondents agreed that extending
center professionals cited the large
the life cycle of storage equipment
amounts of raw materials needed for
is important, emphasizing durability
infrastructure as a key issue.  
to reduce replacement and
3. Physical space constraints. Close maintenance costs. (In contrast, only
to 45.5% of respondents highlighted 15% of the respondents considered
the financial and logistical burden of life cycle extension a top purchasing
limited space. factor for data storage infrastructure
4. Infrastructure costs. High or equipment, and 12% chose
construction costs for sustainable durability as an important factor.)

2 Toward a Systematic Survey for Carbon Neutral Data Centers, IEEE Communications Surveys & Tutorials, Vol. 24, No. 2, Second Quarter, 2022.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 5 |
KEY TAKEAWAYS

These concerns underscore the multifaceted nature of TCO, This is not to say that sustainability does not complicate TCO
integrating both CapEx and operational costs. They also calculations—it can.
directly bear on sustainability. Improving energy efficiency
• For example, power consumption is an inherent part of
reduces both emissions and operational costs—and
the TCO equation, but green energy sources may be more
extending equipment life cycles minimizes e-waste and raw
costly up front. In this instance, the focus on the nature of
material demand.
energy procured adds cost and efficiency constraints to an
already challenging cost efficiency problem.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 6 |
KEY TAKEAWAYS

A Threefold Choice
As the increasing use of AI drives up data When asked about top three barriers
1. scale up (add more
creation, organizations need more room to sustainable data center operations,
for ever-expanding data volumes, and respondents pointed to lack of physical
storage within existing
many struggle with space limitations. space (45.5% of respondents); cost of center footprint)
Spatial constraints affect their ability to constructing storage infrastructure
expand and implement more sustainable (28.5%); and cost of acquiring data
2. scale out (add a
storage solutions. In addition, high up- center components (27%). building or expand the
front costs associated with sustainable
Because of these barriers, organizations
data center footprint)
infrastructure remain a significant barrier
to progress.
often face a threefold choice. To
3. migrate data to
accommodate skyrocketing volumes of
• About 82.5% of respondents said data, they must choose one of the three:
the cloud
they lack the physical space to build
sustainable data storage infrastructure.

Each of these options comes with TCO

and sustainability trade-offs.

Sustainability strategies. The survey

revealed that businesses are increasingly
adopting a number of strategies that
align sustainability with TCO goals.  

Renewable energy sources. Nearly

62% of respondents reported using
renewable energy sources to power their
data infrastructure.

Renewable energy infrastructure.

Close to 58% of survey takers said their
organizations are investing in renewable
energy infrastructure.  

AI-enabled storage and sustainability

operations. Additionally, 55.5%
of represented organizations are
implementing AI-based systems to
optimize storage operations and
enhance sustainability monitoring.

Despite these efforts, challenges remain.

Many companies struggle with high
infrastructure and component costs, as
well as space limitations.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 7 |
KEY TAKEAWAYS

The Path Forward

Addressing these challenges requires a shift in how the
industry approaches TCO and sustainability. Increasingly,
they go together.  When sustainability goals do complicate
TCO plans, attention needs to be paid across the entire
ecosystem to:

Scope 1: Greenhouse gases (GHG)

that an organization emits from
sources it owns or controls directly.
Scope 2: GHG deriving from an
organization’s purchase of electricity,
steam, heat, or cooling.
Scope 3: GHG emissions originating
from business operations by sources
that are not directly owned or
controlled by an organization.

The AI-amplified impact on data center sustainability

calls for innovative and holistic thinking about data
center practices. Among other things, this means looking
beyond operational carbon emissions (Scope 2 and
Scope 3) to include evaluation of embodied carbon—or,
carbon emitted during upstream extraction, production,
transport, bill of material, manufacturing, packaging, and
distribution stages of a product’s life cycle (Scope 3).
As the boom of AI applications stimulates the growth of
data centers, collaboration and innovation across the
supply chain will be key to achieving a sustainable and
more efficient datasphere.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 8 |
Section 1

Challenges to
Decarbonizing
Data Centers

In today’s AI economy, the rapid AI’s energy demands extend beyond of AI systems contribute to a larger
adoption of smart technologies data storage. The most energy- carbon footprint, intensifying the
is driving a sharp increase in data hungry areas include compute, environmental challenges faced by
storage demand. Nearly all survey networking, and data processing. data centers.
respondents (over 97%) expected
The whole supply chain matters, In response to this new demand,
AI to have a substantial impact on
too. Component manufacturing for data centers have begun to
storage needs.
AI—such as specialized processors prioritize sustainability and
and hardware—is highly energy- decarbonization. Nearly all survey
But data always leaves and resource-intensive, which respondents (close to 94.5%)
a footprint: data centers increases the embodied carbon reported that their companies are
consume energy and of these components. (Embodied working to lower the carbon footprint
generate emissions. carbon refers to the quantity of GHG of their data storage operations.
emissions related to the upstream
The AI-driven rising demand for data extraction, production, transport,
storage directly increases the carbon bill of material, manufacturing,
footprint of data centers. As data packaging, and distribution stages of
volumes grow, so does the energy a product’s life cycle.) Consequently,
required to maintain them. Of course, both the production and operation

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 9 |
The survey found that to reduce
their environmental footprint,
companies are primarily adopting
renewable energy (see Figure 1):
• 61% of respondents reported that
their companies use renewable
energy sources to power data
infrastructure.
• Nearly 58% of respondents said
their companies were building new
renewable energy infrastructure to
power data equipment.
• Close to 42% of the data center
professionals surveyed reported
implementing AI-driven solutions
to better determine their data
storage or operational needs.

Another strategy is cloud migration. remains. It is merely transferred

Nearly half of respondents (over from individual companies to the
47%) believed migrating data to cloud providers, who must manage
cloud systems helps reduce their energy and resource demands. This
company’s carbon footprint. Cloud underscores the need for a holistic
service providers, with their approach to sustainability, where
best-in-class practices and both cloud providers and their
infrastructure, have an advantage in clients work together to minimize the
lowering the environmental impact carbon footprint of data.
of data operations. In transitioning to more sustainable
However, enterprises looking to data storage operations, companies
deploy this strategy need to keep in face significant challenges. They
mind that cloud migration can also include physical space constraints,
be seen as shifting responsibility costs, high energy consumption, and
rather than eliminating impact. While effective assessment. Let us take a
cloud providers may operate more closer look at each of these factors.
efficiently, the environmental burden

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 10 |
Challenge: Space and Cost
The top three barriers to sustainable a “lack of physical space” as the top Given the dilemma of having to
data center operations are the lack obstacle. This barrier was particularly choose between a) increasing
of physical space (45.5%), the cost of pronounced in Japan (60%) as well costs of expansion within the same
constructing storage infrastructure as Taiwan and France (both close to space, b) building a new addition
(28.5%), and the cost of acquiring 57%), where it was the top concern. (also costly, often more so than
data center components (27%) (see Building additional physical spaces the first choice), or c) migrating
Figure 2). for data centers can significantly data to the cloud (which passes
Not surprisingly, physical increase direct (Scope 1) GHG the sustainability burden to cloud’s
constraints—such as space emissions, as it involves extensive supply chain), the survey pointed
limitations for new infrastructure— construction activities that consume to a need for innovative solutions to
ranked as the biggest challenge, resources and rare earth minerals, optimize existing space.
with 45.5% of all respondents citing and release GHG.

In response to a different survey

question, 82.5% of respondents agree or
strongly agree with the statement, “My
company does not have the physical
space to build sustainable storage
infrastructure.”

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 11 |
 The need to accommodate more The high costs of constructing and
servers, storage devices, and maintaining green infrastructure
networking equipment intensifies remain a major barrier to
this challenge of limited space. In implementing sustainable solutions.
urban areas, where space is already These expenditures include the
scarce, finding suitable locations upfront investment in renewable
for new data centers is especially energy and the ongoing expenses of
difficult. Several countries, including sustainable IT infrastructure.
Singapore, Japan, Taiwan, France, The survey projects that an
and Germany, have imposed average of $4.9 billion would be
restrictions on the number of needed to invest in more
and locations where data centers sustainable data storage
can be built. These regulations operations globally (see Figure
aim to balance the benefits of 3). While this figure lacks definitive
increased data center resources budget data, its scale highlights the
with environmental protection and magnitude of the challenge—one
efficient land use, underscoring the that requires significant attention,
need for innovative solutions to meet time, and resources.
rising data storage demand.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 12 |
ADDITIONAL DATA POINTS This estimate also covers security sources, and innovative technologies
• Respondents in the United measures. If we were to transition often comes at a premium. Despite the
States (just under 47%), South to 100% green IT products, these potential for long-term savings and
Korea (40%), China, and Australia costs would increase by about environmental benefits, these higher
(both nearly 37%) identified “cost €60,000 to €65,000, which is costs can deter data center operators
of constructing data storage approximately a 10% to 11% increase from making the switch. The need for
infrastructure” as their most
in our budget.” continuous upgrades to keep pace
significant barrier preventing them
Device acquisition costs are the with technological advancements
from building or implementing
largest component of TCO at data further adds to the financial burden.
sustainable data storage solutions.
centers. The initial CapEx required to In the words of one interviewed data
• Breaking down some of the
purchase servers, storage systems, center expert, “shareholders are
common data operating costs
and networking equipment represents typically focused on minimizing
for companies, a German data
a significant share of the overall costs, which can make it hard to
center expert observed, “This
investment—especially at scale, where justify the large expenditures required
includes expenses for the chief
hardware purchases can reach tens of for initiatives like reducing energy
data officer, energy consumption,
millions of dollars. usage or investing in renewable
amortization, and annual
infrastructure investments, totaling Acquiring state-of-the-art, energy- energy sources.”
around €550,000 to €600,000. efficient hardware, renewable energy

Challenge: High Energy Consumption

The continuous operation of data infrastructure (about 49.5%). The high costs associated with
centers consumes large amounts sourcing renewable energy can be a
3. The substantial energy
of energy, contributing significantly significant barrier.
demands of data center
to indirect GHG emissions from
operations (nearly 35%). • Over 61% of respondents
electricity, steam, heat, or cooling
indicated that they are using
(Scope 2). The “always-on” nature Experts warn that without immediate
renewable energy sources as
of data center infrastructure, action, the environmental impact of
part of their mix to power their
including servers and cloud systems, data operations will continue to grow.
data infrastructure, and this has
complicates energy reduction efforts Compounding this issue, many the most significant impact on
and requires innovative solutions. companies reported that they reducing their indirect GHG
In this multi-market research, search out renewable energy, but emissions.
respondents identified their top discover that they have no ability to
• However, renewable energy can
three environmental concerns: choose sustainable energy sources.
be more costly due to the high
1. High energy consumption • Almost 80% of respondents said initial investment in infrastructure,
leading to carbon emissions their company had no option for the need for reliable energy
(roughly 53.5% of respondents). a more sustainable electricity storage solutions, and the
2. Large amounts of raw materials source because it was not complex modifications required
required for data center available in their area. to integrate with existing systems.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 13 |
Challenge: Measurement and Assessment
While data storage itself is a Calling for more real-time sustainability effectively.”
relatively small contributor to information, one of the interviewed A sustainability expert from the
overall data center operations (with US data center experts highlighted Asia-Pacific region reiterated the
compute and processing the need to know how much power importance of transparency in
workloads demanding much more we are using, the fluctuations, and reporting: “Transparent reporting can
energy), the survey asked about any other relevant details: “This help a lot…such as comprehensive
perceptions and measurement information should be logged so environmental impact reports, life
of the environmental impact of we can see what’s happening over cycle assessment, and case studies
data storage. time. Data centers need to see
of successful implementation. Then
• Most respondents (close to 61%) the importance of providing this
there are tools for calculating the
said their companies already information, and companies need to
potential benefits and cost saving of
measure the environmental demand it…real-time monitoring and
sustainable solutions.”
impact of their data storage reporting are essential for managing
through “regular reports detailing
the environmental footprints of
data storage operations.”
• Half of all respondents (over
50%) also said their companies
are using “in-house software to
monitor the environmental
impact of data storage
operations,” while a third (nearly
33.5%) said their companies had
outsourced measurement to
external agencies.
• However, tracking remains
inconsistent, as about 32.5%
of respondents were unsure
or unaware of whether their
companies track these metrics.
There is also a notable lack of
tools and procedures to assess
the environmental impact of data
storage practices. Without proper
assessment methods, businesses
struggle to evaluate the
effectiveness of their sustainability
efforts and make informed decisions
about future investments in
sustainable practices.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 14 |
Section 2

Solutions That
Reduce Data Centers’
Carbon Footprint

The survey makes clear that most data centers could benefit from a reevaluation of traditional data management
practices. While many organizations recognize sustainability’s importance, they often take isolated measures instead
of a holistic approach.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 15 |
 The Importance of Life Cycle
Management, or Lack of It
Ironically, while at least 92% of respondents agreed that extending the life
cycle of storage equipment and infrastructure significantly impacts data
center sustainability, only 15.5% considered life cycle extension as a key factor
in purchasing decisions.
Highlighting the short-term focus of such decisions, a data center expert
from Germany observed:
“Unfortunately, many consumers often choose the cheapest
option without considering the long-term implications. They
might buy a cheaper product only to find it fails after two or three
years, necessitating a new purchase. This cycle of constant
replacement isn’t sustainable. Not enough people think about
this cycle. When we purchase new products, it means more
resources are needed—more components mined from places
like Africa, more manufacturing processes, and more labor. This
impacts the whole ecosystem. The current approach isn’t as
effective as I’d like it to be.”
Extending the life cycle of data center solutions and infrastructure
significantly reduces e-waste and carbon emissions. By maximizing the use
of existing infrastructure, life cycle extensions help mitigate often-overlooked
upstream and downstream emissions.
This highlights the benefits of considering equipment life cycle in
purchasing decisions. Data center operators can evaluate the life cycle of
all components—including power delivery, data processing, and storage
density—to better understand their impact on overall carbon emissions.
When possible, they should select equipment from providers with built-in
circularity programs.
Circularity programs can provide significant sustainability benefits. They
ensure that solutions such as hard drives are either refurbished, repurposed,
or recycled in an environmentally responsible manner. By adopting circularity
principles, data centers can reduce e-waste and the demand for new raw
materials, while minimizing the environmental impact associated with the
disposal of old devices.
This holistic approach not only advances sustainability goals, but also
promotes the efficient use of resources.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 16 |
Understanding the Role of Embodied Carbon
Another gap identified in the survey was under 35%). Scope 2). Embodied carbon refers to
how environmental impact factors into the quantity of GHG emissions related
• However, only about 3% considered
purchasing decisions: to the upstream extraction, production,
low environmental impact a key
transport, bill of material, manufacturing,
• About 95% of respondents expressed factor in purchasing decisions.
packaging, and distribution stages
concerns about the environmental Maintaining awareness of the embodied of a product’s life cycle. Storage and
impact of data center operations. carbon of equipment and infrastructure infrastructure manufacturers as well as
- The top concerns were carbon (Scope 3) can help data centers other data center suppliers ought to
emissions (nearly 53.5%), use of identify opportunities to reduce their provide information on the embodied
raw materials (close to 49.5%), and overall carbon footprint beyond just carbon of their products.
high energy consumption (just operational emissions (Scope 1 and

SEAGATE POV

Storage Media and Embodied Carbon

The table below compares the embodied carbon of solid-state drives (SSDs), hard KEY INSIGHTS

drives, and tape-based storage, highlighting how storage media choices affect total • SSDs have the highest embodied
data center emissions. Seagate Technology analyzed different capacities, usage carbon, both in total and per TB,
patterns, and lifespans over a five-year period, measuring embodied carbon per making them the most carbon-
device and per terabyte (TB) per year. intensive option among the three
storage media.6

Embodied carbon Embodied Embodied carbon • Hard drives exhibit the least carbon
Storage
by product carbon per TB per TB per year footprint, both in total and on a per-
media
(Kg CO₂) (CO₂/TB) (CO₂/TB/year) TB basis, offering the most carbon-
efficient sustainable storage solution.
SSD3 4,915 160 32
• LTO tapes show moderate
embodied carbon, but the
Hard drive4 29.7 <1 <0.2 technology’s annual impact is higher
than that of hard drives.
LTO (Tape)5 48 2.66 <0.6 This analysis highlights the importance of
selecting the appropriate storage media
These numbers reflect Seagate’s analysis based on the following products: mix for both TCO and sustainability.
• Generic data center SSD: 30.72TB
• Hard drive: Seagate 30TB Mozaic 3+™
• Linear tape-open (LTO) 9 tape drive + 1 media: LTO Ultrium

3 The Dirty Secret of SSDs: Embodied Carbon, HotCarbon Workshop on Sustainable Computer Systems, University of Wisconsin-Madison, University of British
Columbia, 2022.
4 Embodied carbon prediction for Seagate 30TB Mozaic Hard Drive per TB for five-year life cycle
5 Seagate analysis of IBM’s breakdown of the CO2 and other positive sustainability impacts of IBM physical tape and power consumption and cooling requirements;
predictions for 1 LTO 9 drive + 1 LTO 9 Media using total embodied carbon with 5 Years life cycle for all technologies, LTO 9: 37 Watt operating and 18-watt idle power.
6 Since there is no publicly accessible life cycle assessment (LCA) data for any data center SSD product, we calculated the embodied carbon of the 30.72TB SSD
using the available public data from the research “The Dirty Secret of SSDs: Embodied Carbon”. We aim to provide accurate figures for data centers to make informed
decisions; however, the NAND industry does not publish LCA data nor undergo third-party audits as Seagate does for hard drives.
DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 17 |
Section 3

Rethinking
Data Center
Sustainability

The survey found that businesses (over 47% of respondents noted

are pursuing two key strategies to this solution) to adoption of a mix
improve data center sustainability: of storage media types (47%) and
1. Adoption of renewable energy. AI-based systems to streamline
Companies are increasingly data storage operations (42%). As
integrating renewable energy noted earlier, there are tradeoffs
sources to power their data and costs (in both financial and
infrastructure. About 62% sustainability realms) that need
said they are currently using consideration, no matter the
renewable sources of energy to solution that is picked.
power their data infrastructure, Integrating renewable energy
and 58% said they are building sources to power data infrastructure
renewable energy infrastructure can significantly reduce reliance
to power data equipment. on fossil fuels, thereby lowering the
2. Deployment of technology- carbon footprint. But it also presents
based solutions. The several challenges. The initial cost
implementation of tech-based of infrastructure, such as solar
solutions is gaining traction, from panels and wind turbines, can be
data migration to cloud systems prohibitively high.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 18 |
Additionally, the intermittent nature integrating renewable energy with Beyond investing in renewable
of renewable energy sources poses existing data center systems can energy, data centers can focus
a challenge, as solar and wind power require complex modifications and on reducing power consumption
generation can be unpredictable upgrades, making the transition at both the equipment and
and dependent on weather time-consuming. Regulatory infrastructure levels. Lowering power
conditions. This necessitates and logistical barriers add to usage not only cuts operational
reliable energy storage solutions to the complexity, as policies and costs, but also enhances the
ensure a consistent power supply, requirements vary across regions. effectiveness of renewable energy
further increasing costs. Moreover, efforts by reducing overall demand.

SEAGATE POV

Storage Media and Power Consumption

Based on Seagate’s analysis of

Power efficiency
power consumption across three Storage media7 Operating power in watts
(W/TB)
storage media types given their
usage patterns and device-level SSD 20 0.5
lifespan over five years, the table
compares:
Hard drive 9.6 0.32
• Operating power in watts,
which is the total power
consumed by a drive when
LTO (Tape) 37 1.1
actively reading or writing data.
• Power efficiency, measured
as the average power These figures highlight that hard drives offer the most energy-efficient
consumed in watts to write or storage option, and are the greenest storage solution, followed by SSDs, with
read 1 TB of data. LTO 9 tape consuming the most power per unit of storage.

7 Based on Seagate’s analysis of LTO 9 Tape Drive: IBM LTO Ultrium Product Specifications; 30TB Hard Drive: Seagate 30TB Mozaic 3+; 30TB SSD: Micron 6500
ION 30.72TB.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 19 |
 Infrastructure-based solutions can In contrast, advanced infrastructure-
play a pivotal role in improving based solutions like HVAC systems
data center sustainability by offer a more comprehensive
addressing energy consumption, approach by not only providing
The survey also found that many data
carbon emissions, and overall cooling, but also managing heating
centers underutilize infrastructure-
environmental impact. and ventilation. These systems often
based solutions to reduce emissions.
Traditional data center cooling incorporate advanced technologies
• Only 23% of respondents use methods, such as computer room like variable speed drives,
HVAC (heating, ventilation, and air conditioning and computer room economizers, and energy recovery
air conditioning) systems to cool air handler units, focus solely on ventilators, which optimize energy
their data infrastructure. cooling the air within the data center use and reduce the overall carbon
to maintain optimal temperatures for footprint.
• Nearly 16.5% had adopted liquid/
equipment. While effective, they
immersion systems for cooling.
can be energy-intensive and
• A small minority (11.5%) reuse heat less efficient in managing overall
generated by data infrastructure. environmental conditions.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 20 |
Section 4

Turning
Challenges Into
Opportunities

To make measurable progress in life cycle of each component allows

sustainability, businesses can benefit In the Shorter Term data centers to identify opportunities
from adopting both immediate and to extend their use and reduce their
long-term strategies. Data centers should develop a environmental footprint. Extending
holistic strategy that goes beyond the life cycle of storage equipment
In the more immediate, short-
adopting renewable energy sources. through regular maintenance,
to-medium term, a more holistic
While integrating renewable energy upgrades, and refurbishment
approach is needed—extending
is a foundational step (though it can programs is crucial. This approach
hardware life cycles, optimizing
be tricky cost-wise), it alone is not cuts e-waste, minimizes the need
energy efficiency, and implementing
for new raw materials, and reduces
circularity programs. enough to address the multifaceted
the energy-intensive processes
challenges of sustainability.
Long-term solutions require associated with manufacturing new
industry-wide, across-the- First, data centers may benefit from equipment. Procurement policies
supply-chain collaboration, implementing regular assessments can prioritize durable, high-quality
financial incentives, and AI-driven and monitoring of the life cycle of components with longer lifespans,
sustainability innovations to ensure all components, including servers, ensuring that sustainability is a key
continued reduction of data centers’ storage devices, and networking criterion in purchasing decisions.
environmental footprint. equipment. Understanding the full

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 21 |
Second, reducing power
consumption can be a critical
part of sustainability efforts. Data
centers can invest in energy-
efficient technologies, such as
liquid/immersion cooling and
HVAC systems, which are more
efficient at dissipating heat. These
systems can significantly lower
energy consumption compared to
traditional cooling methods, leading
to reduced operational costs and
carbon emissions.
Integrating renewable energy efficient, sustainability-promoting new raw materials and minimize the
sources, like solar and wind power, solutions. (For an example from the environmental impact of disposal.
into data center operations can world of data storage, see Areal Minimizing e-waste through
further decrease reliance on Density Innovation and Data Center reuse and repurposing not only
fossil fuels and lower the overall Sustainability on page 23.) supports sustainability goals but
carbon footprint. Optimizing Lastly, establishing a circularity also promotes the efficient use of
energy management practices—for program for the disposal of data resources.
example, by implementing AI-based center components can be valuable. By focusing on these areas, data
systems to monitor and fine-tune Such a program can ensure that centers can significantly enhance
energy use—can also contribute components are refurbished, their sustainability efforts and
to more efficient and sustainable repurposed, or recycled responsibly. work toward reducing their overall
operations. Developing processes to recover environmental impact.
Third, data centers should prioritize valuable materials from end-of-life
the procurement of innovative, cost- equipment can reduce demand for

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 22 |
SEAGATE POV

Areal Density Innovation and Data Center Sustainability

Hard drives are the greenest storage Areal density, the measurement The advancement directly addresses
medium.8 Although hard drives of data stored per unit of surface several sustainability challenges
contribute relatively little to data area on any medium, in a hard flagged by the survey. By increasing
centers’ overall environmental drive specifically refers to the storage capacity without requiring
burden, innovations in hard drive data capacity of a single disk. additional physical space, higher
recording technology can help both Areal density innovation directly areal density helps alleviate space
influences storage efficiency within constraints in data centers, a barrier
optimize data centers’ TCO and
limited physical space and enables cited by 45.5% of respondents.
improve sustainability.
businesses to increase their data Additionally, the reduction in power
With unrivaled economies of scale, storage capacity in the same space consumption it enables aligns with
areal density innovation has become using fewer materials and less power, efforts to lower energy usage and
critical to both AI innovation and effectively reducing their carbon carbon emissions, a primary concern
sustainable data centers. emissions per unit of data storage. for about 53.5% of respondents.

Scaling While Lowering Total Global Energy

and Carbon Impact
For data centers managing by 20289, the proportion of total
large-scale workloads, replacing data center power consumed by
legacy storage devices with next- hard drives in data centers around
generation HAMR (heat-assisted the globe can drop from ~8% to
magnetic recording) drives can lead ~3%. That is a 62% total reduction
to a net decrease in total power in energy consumption by storage
use. At scale, these improvements infrastructure.
translate into meaningful reductions This represents a fundamental
in absolute energy consumption and shift: not only is storage becoming
operational carbon footprint, even as more efficient, but it is also directly
storage infrastructure grows. enabling a net reduction in total
On a global level, this effect energy use and carbon emissions
compounds across the datasphere. across data centers worldwide, even
The transition to higher areal as data retention increases. It’s great
density drives means that, despite news on both data center TCO and
the projected 167% increase in sustainability fronts.
total cloud data storage demand

8 As measured by examining embodied carbon and power consumption.

9 IDC Cloud Infrastructure Index, IDC, 2024.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 23 |
These large-scale benefits are reinforced by efficiency Now, consider how these benefits multiply as HAMR-
improvements at the device level. Based on Seagate’s enabled hard drive technology increases capacities to
analysis of power consumption across multiple 40TB and 50TB.
storage media types (see Storage Media and Power • Imagine the same drive—but with each disk increased
Consumption on page 19), hard drives provide the most to 4TB. The same 10 disks now provide 40TB, a more
energy-efficient, high-capacity storage option. Drive- than 30% increase in capacity in the same physical
level power efficiency factors—including operating power space.
in watts (total power consumed when actively reading or
• With 5TB per disk areal density advancements on the
writing data) and watts per TB efficiency—demonstrate
horizon, the same 10-disk drive will soon store 50TB,
how denser capacities reduce energy consumed per unit
and beyond, further multiplying data center capacity
of stored data.
and power benefits.
By integrating storage solutions with higher areal density,
data centers can expand capacity without requiring
additional space, reduce total energy usage, and
cut operational carbon impact, all while meeting the
increasing demands of AI-driven workloads and cloud
expansion.
For storage at scale, the impact of increased areal density
is profound. Higher areal density allows data centers to
significantly increase their capacity without expanding
their physical footprint, and to achieve sustainability
goals by minimizing the use of space, energy, and natural
resources.
Imagine a 10-disk hard drive, with each disk (also known
as platter) holding 3TB of data. In total, the drive stores
30TB. Consider a data center upgrading each hard drive
in its fleet from 10TB to 30TB, a realistic modern-day
upgrade scenario. This leap in areal density enables:
• Three times or more storage capacity in the same
data center footprint10
• More than a 60% power reduction per TB
• More than a 70% reduction in embodied carbon
per TB11
• 25% lower cost per TB12

10 Comparing Seagate Exos X10 to Seagate Exos X 30TB Mozaic drive

11 Comparing Seagate Exos X10 to Seagate Exos X 30TB Mozaic drive
12 IDC Worldwide 1Q24 HDD Shipments and 4-Quarter Outlook by HDD Segment, IDC, 2024

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 24 |
In the Long Term
In the long term, businesses should its efficiency will be critical. By to scaling operations and storage
explore and leverage AI across their focusing on power-efficient resource infrastructure, data center
operations to curb carbon emissions, planning, optimized training, and life professionals worldwide identified
demonstrating technology’s role to cycle management, AI innovation what they believe to be the top three
drive sustainability. The survey found can be integrated with sustainable motivators:
that 55.5% of respondents believed practices, ensuring its benefits do
• Over 56% cited financial benefits
AI can optimize data storage not come at the expense of the
for adopting greener solutions.
operations to enhance sustainability. environment.
While AI offers significant potential to • More than 40% pointed to tax
Part of longer-term sustainability
drive sustainability efforts, it is crucial relief for adopting greener
planning is the work of changing
to recognize that, if not developed solutions.
minds and culture. It is much easier
with efficiency in mind, AI can also to secure leadership’s buy-in when • At least 38% highlighted the
increase power consumption and sustainability measures enable value of official green company
GHG emissions. greater efficiency, and vice versa. certifications.
To balance AI’s potential with To encourage business leaders to
sustainability goals, optimizing consider sustainable approaches

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 25 |
An expert from the US elaborated on the financial benefits:
“Companies might not want to outlay (a large sum of money) for
a solar array, but they might want to do it and just can’t because
their shareholders would never go for that. If it was basically a net
zero for everybody, it’s a no-brainer. From a federal standpoint, I
got a 30% tax credit on my solar array; it still cost a lot, but I got
30% tax credit back, which offset that cost a little bit.”
Experts emphasized the importance of financial incentives like tax relief
and grants from governments or regional trade blocs like ASEAN and the
European Union.
For example, Germany’s KfW 45 standards specify target energy
consumption levels for buildings, with lower numbers indicating higher
environmental efficiency. Meeting these standards qualifies businesses for
tax relief and government bonuses, encouraging adherence to practices
that lead to more sustainable outcomes.
Collaboration provides businesses with additional avenues to reduce
emissions. Data center professionals saw value in partnering with
governments (around 22%) and other agencies (over 15%) to develop more
sustainable data storage infrastructure.
Data centers may find value in collaborating with industry partners to
develop standardized tools and procedures for assessing the environmental
impact of data storage practices. Such collaboration can help establish
industry-wide benchmarks and best practices, enabling companies to
evaluate performance and identify areas for improvement more effectively.

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 26 |
 Conclusion
The Path to a Sustainable Datasphere

The findings of this research highlight are also tools that can provide insights
the role of industry-wide collaboration for managing sustainability in data
in supporting a sustainable datasphere. centers. Providing comprehensive
By working together, data centers, environmental impact reports,
technology providers, component conducting life cycle assessments,
suppliers, and regulatory bodies can and sharing case studies of successful
develop and implement innovative sustainability interventions can
solutions that reduce energy foster accountability and continuous
consumption, minimize carbon improvement throughout the supply
emissions, and promote the efficient chain. Financial incentives, tax relief,
use of resources. This includes and official green certifications can
adopting advanced cooling solutions, further encourage data centers to
integrating renewable energy sources, adopt more sustainable practices.
extending the life cycle of storage By adopting a comprehensive
equipment, choosing suppliers based approach to sustainability and engaging
on their sustainability commitments, industry stakeholders, companies
and implementing circularity programs can balance the growing demands
for the responsible disposal of of AI and data center operations with
components. Innovation by suppliers environmental considerations.
can play a key role in paving the way Through collaboration and innovation,
to more sustainable and efficient data data centers can help create a
centers. datasphere that is scalable, efficient,
Transparent reporting and real-time and sustainable.
monitoring of environmental impact

DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 27 |
seagate.com
© 2025 Seagate Technology LLC. All rights reserved. Seagate, Seagate Technology, and the Spiral logo are registered trademarks of Seagate Technology LLC in the
United States and/or other countries. Mozaic 3+ is either a trademark or registered trademark of Seagate Technology LLC or one of its affiliated companies in the
United States and/or other countries. All other trademarks or registered trademarks are the property of their respective owners. When referring to drive capacity,
one gigabyte, or GB, equals one billion bytes and one terabyte, or TB, equals one trillion bytes. Your computer’s operating system may use a different standard of
measurement and report a lower capacity. In addition, some of the listed capacity is used for formatting and other functions, and thus will not be available for data
storage. Seagate reserves the right to change, without notice, product offerings or specifications. SC100.1-2503US
DECARBONIZING DATA: STRATEGIC CONSIDERATIONS FOR DATA CENTER SUSTAINABILIT Y IN THE AI ERA | 28 |