The Hidden Cost of Intelligence: Visualizing AI Infrastructure's Environmental Impact on Indian Cities

It started with an Instagram post I couldn't stop thinking about. Someone had shared data about Mumbai's data centers - how they consume as much electricity as thousands of households, how diesel generators near slum areas pump out toxic fumes. The next day, I saw Elon Musk's announcement about building AI factories on the Moon, arguing Earth doesn't have enough energy to scale AI. Something about that felt backwards. We're talking about launching factories into space while people in Mumbai are dealing with power shortages and air pollution from data centers already here? So I started investigating. What began as curiosity turned into weeks of mapping data centers, downloading air quality datasets, and pulling satellite imagery. I wanted to see what AI infrastructure actually costs us. What I found in Navi Mumbai's Airoli area was stark. Forty-seven data center facilities operate near Cheetah Camp , a settlement with crumbling infrastructure, built after residents were evicted decades ago. People report their electricity bills climbing while data centers expand next door, powered by coal plants authorities keep running for 24/7 AI workloads.

What This Talk Covers: This talk walks through that investigation, showing how publicly available tools like satellite thermal imagery, CPCB air quality monitors, power grid data can map this pattern across Indian cities. The methodology is straightforward enough that anyone could replicate it, which feels important. I'll present visualizations showing the correlation between data center locations and environmental impacts, use ML models to forecast which cities face grid crises first under current growth, and discuss accountability frameworks: carbon budgets for AI training runs, mandatory energy reporting, AI workload auditing.

Why This Matters: This topic matters because right now, AI's benefits and costs go to completely different people. Tech companies get profits, vulnerable communities get pollution and power cuts. That's not inevitable , it's a choice. This connects to data visualization's role in advocacy and accountability making invisible power structures visible.

Talk Structure : 1.The question: Why space solutions when Earth's infrastructure creates crises? 2.The investigation: Mapping methodology and findings. 3.What accountability looks like: Frameworks and global examples. 4.Tools and call to action.

Intended Audience: Data journalists, civic tech practitioners, environmental advocates, policy analysts, and anyone curious about technology's real costs. No technical background needed.

Key Takeaways:

• Replicable methodology using only public datasets
• Interactive visualizations adaptable for any city
• Understanding of AI workload auditing as policy tool
• Toolkit for community-driven tech accountability

The question that keeps coming back: why are we dreaming about space-based solutions when we haven't figured out how to build AI infrastructure on Earth without making people's lives measurably worse?