RAWCLIC - Consolidated Composition Data - Data Centers

Descrition

This dataset provides consolidated information on the component and material composition of data centers, developed as part of Deliverable 3.1 of the EU-funded RAWCLIC project. It supports quantitative assessment of raw material demand, resource intensity, and environmental footprints associated with the global data center landscape — a critical infrastructure underpinning the digital transition.

Data centers are classified using a three-level scheme covering facility type (Cluster, Colocation, Enterprise, Edge, HPC, Hyperscale, MicroData), size (Small, Medium, Large), and primary application (AI, Cloud, Storage, Traditional, Retail, Wholesale). The physical structure is modeled hierarchically — from facility-level rack counts through functional rack-type distributions (Compute, Storage, GPU/AI, Network), down to individual hardware units and their printed circuit board (PCB) areas. For each parameter, minimum, mode, and maximum values are provided, enabling both deterministic and probabilistic use of the data.

The dataset covers PCB area estimates for all major hardware components (AI/Compute/Storage/Network mainboards, GPU boards, DIMMs, NIC boards, PCI riser cards, PSUs, storage controllers, U.2 boards, drive backplanes), disaggregated by data center type, size, and application. Statistical descriptors include mean, median, mode, standard deviation, and 2.5th/97.5th percentiles, derived from Monte Carlo simulation. Data quality metadata and full source references are provided in the accompanying report (datacenter_comprehensive_report.pdf).

Note on Data Transmission Infrastructure

Data transmission infrastructure is functionally a subset of the broader data center ecosystem. Network hardware — including NIC boards and network mainboards — is an integral component of every rack configuration and is therefore captured within this dataset through the Network rack type. The hardware building blocks common to both domains are represented here, making this dataset directly relevant to assessments of transmission-related raw material demand.

About RAWCLIC

RAWCLIC (Future RAW materials demand, supply and sustainability in the face of CLImate Change) is a 48-month EU-funded research project coordinated by BRGM, with Empa as lead beneficiary for Work Package 3. The project develops knowledge on future raw material demand, supply, and environmental impacts driven by the twin transition in the EU, empowering industry and policymakers to make informed, evidence-based decisions in support of sustainable transition pathways. Further information at www.rawclic.eu.