SoftBank’s €75bn AI bet exposes EU’s tech sovereignty gaps and extractive growth model
Original framing: “SoftBank pledges €75bn to build Europe’s biggest AI facility in France” — Financial Times
The scientific consensus on AI’s environmental impact (e.g., Strubell et al., 2019) highlights that training large models can emit as much CO2 as five cars over their lifetimes, yet the €75bn pledge makes no mention of carbon accounting or renewable energy commitments. SoftBank’s reliance on cloud infrastructure (e.g., NVIDIA GPUs) also ties the project to the semiconductor industry’s water and mineral extraction crises (e.g., cobalt mining in Congo). The lack of peer-reviewed feasibility studies for such a massive facility raises questions about its long-term sustainability.
SoftBank’s €75bn AI pledge is less a triumph of European tech ambition than a symptom of the EU’s structural dependency on foreign capital and its failure to develop alternative models of innovation.
The deal reflects Japan’s post-bubble strategy of leveraging cheap credit to pursue global dominance, but it risks repeating the mistakes of past eras—overcapacity, environmental harm, and social inequality—while masking these risks under the sheen of 'AI leadership.' The absence of indigenous critiques, historical parallels, and marginalized voices in the narrative reveals how the techno-utopian frame serves to naturalize extractive growth, obscuring the power asymmetries between Son’s vision and Europe’s democratic priorities. A systemic response requires not just regulatory oversight but a reimagining of AI as a public good, with energy democracy, data sovereignty, and participatory governance at its core—models already being tested in pockets of the Global South and within Indigenous communities. The real question is whether Europe will repeat the errors of Silicon Valley’s extractive capitalism or chart a new path, one where technology serves the many, not the few.