Registration form
Attendance is free and open to everyone interested. Please register via the link above, and you will receive the Zoom meeting details one day before the seminar.
Julia Hess & Maria Nowicka, interface
As the EU prepares to significantly expand its compute infrastructure and chip production, critical questions about the environmental implications of this build-out remain surprisingly rarely debated. Current initiatives include plans to potentially triple data-centre capacity by 2030 under the forthcoming Cloud and AI Development Act, establish five AI Gigafactories, and further scale semiconductor production under a proposed Chips Act 2.0.
This talk argues that the question of whether Europe is building the right infrastructure for its actual needs is closely linked to environmental sustainability. Expanding compute capacity is not only a matter of scale, but also of purpose and design.
Drawing on recent publications from interface’s Global Chip Dynamics Programme, the presentation examines the environmental implications of the AI-driven build out of compute infrastructure. Particular attention is given to the climate and resource impacts of semiconductor manufacturing. The talk also explores a possible link between the recent rise in emissions from chip production and the AI boom. Energy represents another key dimension and an emerging bottleneck. Electricity demand from data centres is rising rapidly. This growth is closely tied to decisions about the purpose large-scale compute infrastructure is intended to serve – such as training or inference – and the hardware requirements resulting from that.
The talk concludes by connecting policy initiatives that are often treated in isolation. It highlights the second-order effects emerging across AI, semiconductor, and energy policy that remain largely absent from current debates.
This talk argues that the question of whether Europe is building the right infrastructure for its actual needs is closely linked to environmental sustainability. Expanding compute capacity is not only a matter of scale, but also of purpose and design.
Drawing on recent publications from interface’s Global Chip Dynamics Programme, the presentation examines the environmental implications of the AI-driven build out of compute infrastructure. Particular attention is given to the climate and resource impacts of semiconductor manufacturing. The talk also explores a possible link between the recent rise in emissions from chip production and the AI boom. Energy represents another key dimension and an emerging bottleneck. Electricity demand from data centres is rising rapidly. This growth is closely tied to decisions about the purpose large-scale compute infrastructure is intended to serve – such as training or inference – and the hardware requirements resulting from that.
The talk concludes by connecting policy initiatives that are often treated in isolation. It highlights the second-order effects emerging across AI, semiconductor, and energy policy that remain largely absent from current debates.