As the rapid rise of AI technologies continues to fuel global innovation, the environmental impact of cloud computing is becoming a critical concern. Columbia Engineering’s Climate Week opened with a vital conversation on how to mitigate the carbon footprint of the ever-growing computing infrastructure. The event, "Sustainable Cloud Computing and AI," featured prominent voices from academia and industry to tackle this pressing issue, including Associate Professor Asaf Cidon of Columbia’s Electrical Engineering Department.
In his opening remarks, Professor Cidon set the stage by highlighting the scale of the challenge. Cloud datacenters already account for approximately 2 percent of global electricity consumption, and with the exponential growth in AI, particularly with the explosion of generative AI, this number could rise even more steeply. Cidon says that we need to rethink everything from the chips that power cloud servers to the algorithms that drive AI models.
Cidon’s research at Columbia focuses on the intersection of cloud infrastructure and energy efficiency, making him a key voice in the ongoing discussions about sustainability in computing. The panel, part of Columbia Engineering's larger Climate Week efforts, convened experts from IBM, University of Michigan, Meta Fundamental AI Research, Cornell Tech, and University of Pennsylvania, and the Columbia’s Data Science Institute, who discussed innovations aimed at creating carbon-aware cloud systems, energy-efficient hardware, and software that can minimize the environmental impact while continuing to support AI's expansion.