The rapid rise of AI technologies generates unprecedented data volumes, leading to an explosion in storage demand and computing power, which intensifies the energy consumption of data centers and their carbon footprint. Seagate's report on data decarbonization calls for rethinking the architecture of data centers not solely from a performance perspective, but by rigorously balancing economic costs and carbon footprint.
According to Goldman Sachs, by 2030, global energy demand from data centers is expected to increase by 165% compared to 2023. Seagate's report, a major player in data storage, based on a global study conducted by the independent firm Dynata and the consultancy Current Global, sheds light on the environmental challenges related to digital infrastructures.
Energy consumption is now a major concern for 53.5% of decision-makers surveyed, forced to juggle between infrastructure expansion, cost control, and reducing environmental impact. The survey highlights several key insights:
  • Increasing pressure related to AI: 94.5% of respondents report an increase in their storage needs, and 97% anticipate that the development of artificial intelligence will further intensify this demand;
  • A gap between environmental concern and purchasing decisions: while nearly 95% of respondents express concern about the ecological impact of their data centers, only 3.3% prioritize it as a criterion when purchasing equipment;
  • Structural barriers to sustainability: among the main obstacles identified are energy consumption (53.5%), raw material needs (49.5%), space constraints (45.5%), infrastructure costs (28.5%), and acquisition costs (27%);
  • A disconnect around the equipment lifecycle: although 92.2% of participants recognize the importance of extending the lifespan of storage devices, only 15.5% consider this criterion as decisive in their purchasing decisions. A tension revealing the complexity of the trade-offs at play.
Jason Feist, Senior Vice President of Cloud Marketing at Seagate, emphasizes:
"Data centers are under particular scrutiny, not only because they manage current AI workloads, but because they are becoming one of the most energy-intensive sectors of the digital economy. This requires a fundamental transformation in terms of data infrastructure design. It's not about compromising between cost and sustainability, but about optimizing both aspects."

Towards an Ecological Transformation of Storage

According to IEEE, by 2030, data centers could account for up to 8% of global carbon emissions, compared to 0.3% in 2022. In response to this issue, some companies have initiated a structured response:
  • 61.8% use renewable energy sources;
  • 57.8% invest in green energy infrastructures;
  • 55.5% implement AI-driven management systems to optimize operations and enhance environmental traceability.
But obstacles persist: lack of space, high initial costs, and fragmented approaches still hinder the alignment between economic performance and sustainability.
Seagate proposes an approach based on three strategic pillars.
The first is based on technological innovation. It advocates adopting eco-efficient technologies such as liquid cooling and advanced storage solutions like the Mozaic 3+ platform based on HAMR technology, which reduces the carbon footprint by 70% per terabyte and lowers storage costs.
The second lever concerns extending equipment lifecycle. For Seagate, real-time environmental monitoring and transparent reporting could foster accountability across all data centers.
Finally, the third pillar refers to collective responsibility. Reducing emissions across all scopes (1, 2, and 3) requires close coordination between manufacturers, cloud providers, and end users.
Jason Feist concludes:
"Sustainability is not an issue that can be solved alone. A comprehensive approach including infrastructure, lifecycle management, and accountability at the sector level would ensure that the growth of AI and data center operations does not come at the expense of the environment."

Translated from IA, stockage massif et durabilité : quelle stratégie pour les centres de données ?

To better understand

What is HAMR technology and how does it contribute to reducing the carbon footprint of data centers?

HAMR (Heat-Assisted Magnetic Recording) technology uses heat to increase storage density on hard drives, allowing more data to be stored without increasing the physical size of the disk. This significantly reduces the carbon footprint per terabyte stored by optimizing material resource use and decreasing overall energy consumption in data centers.

What are 'scopes' 1, 2, and 3 in terms of carbon emission reduction, and how are they relevant to data centers?

'Scopes' 1, 2, and 3 are categories of greenhouse gas emissions established by the GHG Protocol. Scope 1 covers direct emissions from sources owned or controlled by the organization. Scope 2 concerns indirect emissions from the generation of purchased electricity. Scope 3 includes all other indirect emissions in the value chain, such as from suppliers and consumers. For data centers, these scopes are crucial for evaluating and reducing total carbon footprint by involving all sector stakeholders.