UN report flags AI data centres’ resource surge; Israeli startups offer fixes

A United Nations University study warns that AI data centres could consume 945 terawatt‑hours of electricity each year by 2030, roughly three times the combined annual electricity use of Pakistan, Bangladesh and Nigeria. The study estimates water use could meet the basic domestic needs of 1.3 billion people and a land footprint of more than 14,500 km², about twice the size of the Jakarta metropolitan area. Over 90% of AI‑specialised computing capacity is concentrated in the United States and China, while more than 150 countries lack significant domestic AI infrastructure, underscoring a widening environmental burden beyond carbon emissions.

In parallel, Israel’s startup ecosystem is mapping around 60 companies that develop energy‑generation, storage, cooling and grid‑management technologies for data centres. The map highlights solutions such as Phinergy’s aluminium‑air generators, Salion Energy’s next‑generation batteries and advanced heat‑reuse cooling systems. As global tech firms invest billions in AI infrastructure, these startups aim to help prevent power‑supply crises and support the “bring‑and‑build your own energy” approach emerging among major corporations.