Why it matters to calculate the carbon footprint of your data
According to IDC projections, 64.2 zettabytes of data were generated or replicated globally in 2020, with a compound annual growth rate of 23% forecast between 2020 and 2025 — over 180 zettabytes expected by 2025. Every company produces and stores a growing volume of documents, databases, emails and backups, and this volume carries an environmental cost as well as an economic one.
Global data centres consume around 416 terawatt-hours of energy per year, with some estimates reaching up to 770 terawatt-hours. Of this consumption, only 11% is directly attributable to powering storage drives: 43% goes to servers and 43% to cooling and power systems. A significant share of the energy used for cooling and processing is nevertheless related to data management.
One often underestimated factor: NRDC and IBM statistics indicate that servers operate on average at only 12-18% of their capacity, while consuming almost the same energy they would at 80%. A large part of the storage capacity already manufactured worldwide therefore remains underused.
Manufacturing the hardware weighs more than using it
Energy consumption during use is not the main item in the carbon footprint of storage. A life cycle analysis of a 1TB Seagate Momentus drive shows that only 16% of the climate impact comes from operational energy consumption. The rest comes from:
- Material extraction and pre-processing: 62%
- Assembly: 11%
- Transport and distribution: 8%
- Disposal and recycling: 3%
Another study (Tannu and Nair, on the embodied carbon of SSDs) estimates that manufacturing 1 TB of HDD capacity generates around 20 kg of CO₂, to which a further 76 kg of CO₂ is added to power that capacity over a five-year useful life. According to the same research, reusing drives is the end-of-life strategy with the greatest environmental benefit: around 5 kg of CO₂ saved for every six months of extended useful life.
It should also be considered that most cloud storage providers replicate data at least 3 times to guarantee durability (the 3-2-1 backup rule), and this factor multiplies further when data must be distributed across several geographical regions. One functional TB of data can therefore require 3 TB (or more) of actual physical capacity, with the related carbon cost.
How Certiblok®'s decentralised storage works
Certiblok® provides S3-compatible cloud object storage that leverages already existing, independently managed drive capacity worldwide, instead of requiring new dedicated capacity to be built. The system works like this:
- Files are encrypted end-to-end (AES-256-GSM) and segmented into 64 MB units
- Each segment is split into 80 fragments through Reed-Solomon erasure coding
- The fragments are distributed across different, geographically dispersed drives
- The file can be reconstructed from just 29 of the 80 fragments
Instead of traditional 3-2-1 replication, this approach produces an effective expansion factor of 2.7x (80/29) rather than 3x or more, while still maintaining durability above 11 "nines" (99.999999999%) — which includes protection against risks affecting entire geographical areas, such as fires, floods or power outages.
Our analysis identifies five main levers through which this model reduces emissions:
- Optimisation of existing capacity: a share of the Certiblok® network comes from drives already in operation, with minimal incremental carbon impact
- Life cycle extension: drives are used beyond the traditional 3-5 year limits, with failures handled through erasure coding
- Reduced replication: a 2.7x factor against the 3x (or more) of traditional approaches
- Optimisation of geographical distribution: fragments are already distributed globally, with no need for dedicated regional copies
- Smart HDD-SSD balancing: since the life cycle carbon cost of an SSD is about twice that of an HDD, reducing the share of SSD cache where it is not needed generates further savings
The numbers: how much you really save
Our calculation model compares the carbon impact of one TB of data stored for 3 years in four scenarios: Hyperscaler, corporate Data Centre, Certiblok® Standard (drives already supplied and powered) and Certiblok® Blended (weighted mix of network profiles).
| Comparison | Traditional approach (kg CO₂/TB, 3 years) | With Certiblok® (kg CO₂/TB, 3 years) | Reduction |
|---|---|---|---|
| vs. Hyperscaler | 362 | 15 (Certiblok® Standard) | -95% |
| vs. corporate Data Centre | 679 | 15 (Certiblok® Standard) | -98% |
| vs. Hyperscaler | 362 | 112 (Certiblok® Blended) | -66% |
| vs. corporate Data Centre | 679 | 112 (Certiblok® Blended) | -83% |
Certiblok® Standard: drives already supplied and powered. Certiblok® Blended: weighted mix of network profiles (Standard, Reused, New Nodes).
To give a practical scale to these numbers: a car produces around 0.24 kg of CO₂ per kilometre travelled. The saving achievable on a single TB of actual storage, in the most conservative scenario, is therefore equivalent to the emissions of around 1,000-2,000 km of car travel.