May 04, 2026

India’s data centre buildout is starting to look like a water story as much as a compute story. In New Delhi, environmentalists are warning that the country’s AI and cloud boom could swallow nearly 37.5 billion litres of water a year—and they’re taking it straight to the Prime Minister’s Office. The government’s response: industry is adopting “advanced cooling technologies,” but the underlying momentum is hard to miss, with capacity up from 375MW (2020) to 1,500MW+ (2025).

The Big Stories

Environmentalists Warn India’s AI Boom Threatens Water Security puts an unusually concrete number on a problem the industry often hand-waves: scale. NatConnect’s letter flags a national water-security risk from AI and cloud growth, while the Ministry of Jal Shakti points to efficiency measures and notes how quickly capacity has quadrupled since 2020. What matters here is less the exact litre estimate than the politics it invites: once water becomes a headline risk at the centre of government, permitting and operating norms can change quickly—especially in places where water is already contested.

In Central Texas, the local politics are already moving in that direction. Advocates: Austin must pair environmental protection with affordability describes how environmental and community leaders are tying growth debates together: green space, displacement, and infrastructure strain. Travis County Commissioner Ann Howard explicitly calls out data centres moving into the region and urges requirements like reclaimed water use or battery operations. The signal for operators and investors is that “sustainability” is being reframed as a hard constraint with local enforcement teeth—not just a reporting exercise—right as Austin’s broader affordability fight sharpens scrutiny of who gets scarce power and water.

Behind the Headlines

Coratia Technologies builds Indian robots to protect undersea infrastructure reads like a defence-tech funding round, but it’s really a connectivity resilience story. An Odisha-based startup has raised ₹22 Cr and landed a ₹66 Cr iDEX contract to supply underwater robots, including a military-grade platform, to the Indian Navy—aimed at inspections, defence, surveying, and undersea infrastructure protection amid submarine cable threats. The bigger point: as subsea cables become a front-page vulnerability, more of the “data centre ecosystem” spend shifts from land (campuses, substations) to the seabed (inspection, monitoring, and response). That’s a quiet but meaningful expansion of what critical infrastructure protection looks like in the AI era.

Students protest Barton Peveril’s rollout of generative AI is about a college, but it’s a useful proxy for the next wave of AI pushback: not against data centres directly, but against institutional AI deployment with weak consent and unclear governance. Barton Peveril rolled out a generative model (“Barton AI”) and a personal-development branch (“Barton PD AI”); after a student-led campaign, the PD element was temporarily suspended. The piece cites broader adoption stats—student AI use jumping from 53% to 88% in 2024–25, and 76% of teachers using AI—which underlines the tension: usage is exploding, but legitimacy is fragile. For the infrastructure world, this is the demand-side counterpart to water and power fights—AI growth won’t just be capped by electrons and litres, but also by whether institutions can deploy AI without triggering backlash that forces policy intervention.