Episode 28 of The AI Podcast with Fexingo. AI data centers consume enormous amounts of water for cooling — and it's becoming a real business constraint. Lucas and Luna dig into a single shocking data point: training GPT-4 probably consumed over 2 million liters of fresh water, enough to fill an Olympic swimming pool. They explore where the water goes, why hyperscalers are facing local backlash, and how companies like Microsoft and Google are trying to square AI growth with water stewardship. Plus, a look at what this means for investors: which chip makers and data center operators are most exposed to water risk, and why some locations are becoming off-limits for new builds. If you care about AI infrastructure, this episode will change how you think about the physical costs of the technology.

#AI #DataCenters #WaterCooling #Sustainability #Microsoft #Google #NVIDIA #SOXX #Hyperscaler #Infrastructure #Technology #Business #FexingoBusiness #BusinessPodcast #AIInfrastructure #ClimateRisk #WaterScarcity #ESG

Keep every episode free: buymeacoffee.com/fexingo