Photo: Alexi Bienvenu © LFDE
By Alexis Bienvenu, Portfolio Manager, La Financière de l’Échiquier (LFDE)
“AI, yes — but not in my backyard”: across the world, opposition is mounting against the construction of new data centers. In the United States alone, more than 300 local moratoriums have been recorded since 2023. Northern Virginia, the world’s leading data center hub, has become a symbol of this resistance. Significant mobilizations have also emerged in Arizona, Texas, Pennsylvania, Michigan and New York State. Europe, India and Brazil are following suit.
Residents are notably denouncing higher electricity prices, the noise generated by these facilities, pressure on water resources and the transformation of landscapes into industrial zones.
The risk is already tangible for investors. According to Morgan Stanley[1], nearly $286 billion of data center projects were delayed or cancelled between 2025 and the first quarter of 2026. The reasons are not limited to local opposition, but it is a contributing factor, leading to construction delays and higher costs.
At the heart of the controversy lies the issue of resources, above all water. New AI-focused computing centers generate enormous amounts of heat, requiring cooling systems that often rely on evaporation. This consumption is particularly critical in a context marked by intensifying droughts and heatwaves. In regions already under water stress, such as the western United States, opponents of data center projects highlight the trade-off between the needs of households, agriculture, industry and natural ecosystems.
Criticism is not confined to water. Data centers are also accused of straining power grids, contributing to greenhouse-gas emissions and sealing off land. Morgan Stanley estimates that the global buildout of AI-related computing infrastructure could generate around 2.5 gigatonnes of cumulative CO₂ emissions by 2030, equivalent to roughly 40% of current annual U.S. emissions.
To mitigate these negative spillovers, a range of solutions is being explored. These include liquid cooling technologies, which are more water-efficient, as well as water recycling and the recovery of waste heat. Reducing the size of data centers would also help improve their social acceptance.
But the intensity of competition among the major players, whether companies or states, leaves little time for such solutions to emerge. Especially since data center construction is also driven by geopolitical urgency. Computing capacity has become critical infrastructure for national or regional sovereignty, on par with power grids or rare-earth mines. The rivalry between the United States and China in this field is fueling a race for digital capacity.
Against this backdrop, AI investors face a choice between two paths. The first is to fully incorporate resource constraints and social acceptability into the AI factory model, at the cost of likely lower short-term profitability for hyperscalers such as Alphabet, Amazon, Microsoft and Meta, but with greater resilience. The second, simpler but less sustainable and therefore more fragile in the longer term, is to shift investment toward regions that are currently less demanding, such as Malaysia, the United Arab Emirates, Saudi Arabia or Indonesia. But there is no guarantee that this availability will last: these territories may seek to monetise, in one form or another, their greater tolerance for the exploitation of local resources.
Taking ESG factors into account in the AI economy is therefore no longer merely a matter of social responsibility; it is becoming a determinant of long-term valuation. Building computing centers on the Moon or on Mars would not change that: resources there would remain limited and opposition inevitable. The backlash against data centers is a useful reminder that the AI economy is not only virtual, but also deeply material and political.
Disclaimers : Ces données et opinions sont fournies à titre d’information et, de ce fait, ne constituent ni une offre d’achat ou de vente d’un titre, ni un conseil en investissement ni une analyse financière. Les opinions sont celles de l’auteur, elles ne sauraient en aucun cas engager la responsabilité de LFDE. Les performances passées ne préjugent pas des performances futures.
[1] Morgan Stanley Research : “Power Struggle: AI, Data Centers & Local Backlash”, 14/07/2026
Retrouvez l’ensemble de nos articles Strategic Case
