Digital technology, metals, and biodiversity: science for decision-making

Metals: an emerging pressure on ecosystems

The transition to a carbon-free economy depends on metals: they are everywhere, in batteries, solar panels, wind turbines, electric vehicles, cables, servers, smartphones, etc. But this material transition is creating new pressure on ecosystems.

Order of magnitude

Mining uses significant amounts of land and water, releases heavy metals and other pollutants, and permanently alters landscapes. In some countries, it adds to already significant ecological vulnerabilities.

The CO2 emissions of various base metals are shown in Figure 1:

More broadly, a life cycle assessmentconducted by ADEME on all digital uses in France shows that, on a national scale, digital equipment and infrastructure account for 10% of France’s electricity consumption and 2.5% of its carbon footprint.

Companies engaged in the digital transition are therefore faced with a key question: how can they reduce their impact on biodiversity when the value chain is global and data is incomplete?

The answer requires scientific rigor, appropriate measurement tools, and a detailed understanding of the interactions between human activities and natural environments.

Digital technology and AI: a very real footprint

As artificial intelligence establishes itself as a driver of digital transformation, its environmental footprint is becoming a major topic of debate.
AI models, often perceived as “intangible,” nevertheless require significant physical resources: servers, graphics cards, cooling infrastructure—and therefore metals, energy, and water.

🔍 What Mistral AI’s recent study shows

In 2025, Mistral AI published a comprehensive analysis of the environmental footprint of its Mistral Large 2 language model, adopting a full life cycle approach (training + inference).
Some key figures from this study:

• Greenhouse gas (GHG) emissions: 20.4 kilotons of CO₂ equivalent for training, over 18 months of use;
• Water consumption: 281,000 m³;
• Depletion of abiotic resources: 660 kg antimony equivalent (Sb eq), an indicator of the use of non-renewable resources—including metals and minerals used in the manufacture of servers and electronic components;
• Inference (a query of approximately 400 tokens, via “Le Chat”): 1.14 g CO₂e, 45 mL of water, 0.16 mg Sb eq per query.

These figures represent a significant step forward: for the first time, a major AI player has published a comprehensive environmental assessment that includes not only CO₂ emissions but also water consumption and abiotic resources depletion —whereas most analyses are limited to energy.

⚠️ But this study remains incomplete

The “abiotic resource depletion” (ADP) indicator does include metals, but without specifying which metals, where they are extracted, or which ecosystems are affected.
The study therefore does not yet link these material flows to biodiversity—an essential link in understanding the real impact of digital technology on living organisms.

✔️ What this reveals

This publication illustrates a fundamental trend:
digital players—including those in AI—are beginning to measure their environmental impact using a life cycle approach.
But the “metals and biodiversity” dimension remains largely undervalued.

This is precisely where Sayari brings unique value: by linking material flows (copper, cobalt, aluminum, lithium, etc.) to their effects on ecosystems, and by providing science-based decision-making tools.

Measure to act: the Sayari methode

For more than a decade, Sayari has been linking scientific knowledge and operational decision-making.
Our teams, experts in life cycle assessment (LCA) and biodiversity impact assessment, have developed a specific method for assessing the impact of metal sourcing.

This method is based on three key principles illustrated in Figure 2:

1. A comprehensive life cycle approach: It takes into account the entire process, from extraction to secondary production (recycling), to assess the Biodiversity Impact Score (BIS) of each metal.
2. Weighting by ecosystem vulnerability: The impact of pollution or land use change depends on where it occurs. Our method incorporates this variability by cross-referencing the pressures exerted and local ecological sensitivity.
3. Data quality assessment: Each result is accompanied by a confidence index to encourage transparency and continuous improvement in data quality.

This approach enables companies to identify priority areas for action, compare supply chains (primary vs. recycled), and integrate biodiversity into their purchasing, eco-design, and CSR strategy decisions.

A concrete example: collaboration with L’Oréal

In 2024, Sayari applied this method to an industrial case: assessing the biodiversity impact of L’Oréal’s beauty devices.
The goal was to understand how the materials used—particularly aluminum, copper, and cobalt—affect ecosystems, and how alternatives or recycled materials can reduce these impacts.

To illustrate this point, we will look at the results from a cobalt mine in Morocco, which was studied outside of the collaboration with L’Oréal.

This collaboration has enabled L’Oréal to (see Figure 4):
• Prioritze its actions on the most sensitive materials;
• Compare production chains;
• Integrate biodiversity into its performance indicators, alongside climate and resources.

This work illustrates a shared conviction: biodiversity is becoming a criterion in industrial decisions, just like carbon.

Towards a science-informed economy

Although most current metal sourcing initiatives focus on reducing CO2 emissions, a few proposals are being developed to address biodiversity issues. These include:

• International Council on Mining and Metals (ICMM): the “Achieving No Net Loss or Net Gain of Biodiversity” project consists of assessing and addressing significant risks and impacts on biodiversity and ecosystem services in order to avoid and minimize impacts, restore affected areas, and finally offset residual impacts to achieve a minimum of net loss/net gain of biodiversity at closure of the mining facility.
• Responsible Mining Foundation (RMF): establishment of a Responsible Mining Index (RMI) until 2022, which assessed the performance of large mining companies using indicators related to biodiversity and ecosystems, among others.
• WWF: The report “Critical Minerals at a Critical Moment” published in May 2025 assesses the risk that increased demand for minerals poses to nature and provides policy recommendations.

However, assessing the biodiversity impact of metal sourcing is only the first step.
The method developed by Sayari is part of an international dynamic:

• Contribution to the European Commission’s future PEF (Product Environmental Footprint) method;
• Participation in the United Nations Life Cycle Initiative;
• Collaborations with industrial sectors, NGOs, and public institutions.

Our ambition: to enable every stakeholder—businesses, local authorities, decision-makers—to base their choices on scientific knowledge. Because only an approach based on measurement and rigor can guarantee truly sustainable decisions.

Science for decision making 

The digital transition and the ecological transition can no longer be considered separately.
The technologies that are transforming our lives will gradually become rooted in respect for life.

At Sayari, we place science at the heart of decision-making. So that every piece of data, every method, every assessment contributes to a future where technological progress and biodiversity no longer conflict, but evolve together.

Bibliographic sources

RÉSUMÉ À L’INTENTION DES DÉCIDEURS DU RAPPORT DE L’ÉVALUATION MONDIALE DE L’IPBES DE LA BIODIVERSITÉ ET DES SERVICES ÉCOSYSTÉMIQUES – IPBES, 2019

Haut-commissariat à la stratégie et au plan – Note d’analyse 96, Comment évaluer l’externalité carbone des métaux, 2020

Mistral AI – Notre contribution pour la création d’un standard environnemental mondial pour l’IA, https://mistral.ai/fr/news/our-contribution-to-a-global-environmental-standard-for-ai consulted on 20/10/2025

ICMM – Achieving No Net Loss or Net Gain of Biodiversity – Good Practice Guide, March 2025

Responsible Mining Foundation – RMI Report 2022

WWF – Critical Minerals at a Critical Moment, May 2025