Home › Hard problems › Chips, compute and energy

Chips, compute and energy

Power and cooling for AI data centres

Solo: possibleOne person can test a narrow slice of it.

the gridyearsa new site waitsfor powerone rack drawsover 100 kWcooling and water can't keep up
AI racks now draw over 100 kW each, grids and cooling cannot keep up, and new data centres wait years for power.

Why it is still open

Racks now draw over 100 kW, and grids and cooling can't keep up. New sites wait years for power.

An experiment one person can run in 8 weeks

Model the power, water and cost of a 10 MW AI data centre in 3 Indian cities from public data.

Verdict: Solo: possible

A strong explainer built from public data.

The solo test behind the verdict

  1. Can it be measured? There is a number that shows you did better, such as an error rate, a catch rate or a cost per answer.
  2. Is it cheap to try? It runs on a laptop, free open models and open data, or under about ₹50,000 of rented GPU time.
  3. Is there a public yardstick? A benchmark or dataset others use, so your result can be compared and believed.
  4. Would a small win get noticed? A lab, a funder or a customer would care about a 10% improvement, not only a full solution.

Four yeses means Solo: strong. Two or three means Solo: possible. Fewer means Follow, don't build.

PreviousConnecting thousands of chipsNextAI for chip design

All 79 hard problems · Working on this, or know of a result we missed? basant@cleandataengineering.ai