Sustainability company Ecolab has highlighted the continued and increasing demands placed on natural resources by AI computing technology, particularly the use of water.
It aims to balance the protection of water resources with data centre profitability, using the same machine learning-based technologies in the sector’s own operations.
Water is a resource under increasing demand due to the additional power requirements of digital technology operations.
Water is used by data centres primarily for cooling to prevent hardware overheating during intensive computation. The cooling of data centres can consume up to 40% of a facility’s power, although that figure depends on where data centres are located. In line with the growth in demand for power, AI’s use is expected to consume as much water annually as the United States’ total drinking water supply.
According to Ecolab forecasts, by 2030, AI ecosystems (typically based in data centres) may require as much electricity annually as India uses in a year.
Companies that use AI intensively (or any other high-performance computing (HPC) facilities) can chose data centre providers in colder climates, and will often benefit from lower costs and local green energy, such as available in Scandinavia and Iceland. The pay-off is slightly slower access times and data movement restrictions (latency and throughput).
Yet AI holds the promise of improving innovation and efficiency in circular water management – essentially using the technology to solve one of the problems of its own making.
Ecolab specialise in water management, with a self-proclaimed mission to protect water and simultaneously help businesses to grow. Christophe Beck, Chairman and CEO of Ecolab, spoke of the opportunities associated with developments in AI. “It offers an unprecedented opportunity to address the water crisis through proprietary connected systems and […] insights,” he said in a press release
“By using AI to reduce, reuse and re-purpose water throughout business operations, we can continue to win the right way […] while protecting natural resources.”
The company says it’s helped customers conserve more than 226 billion gallons of water, equal to the drinking needs of 781 million people. Ecolab says it aims to “conserve 300 billion gallons of water annually, equivalent to the drinking water needs of 1 billion people” by 2030.
Ecolab has formed a partnership with Digital Realty, a real estate investment trust company, to implement an AI-powered water conservation system in 35 of Digital Realty’s data centres in the US.
It’s hoped that such partnerships will allow organisations to use data centres close to the worldwide ‘hubs’ where connectivity is best (in Europe, for example, the FLAP group – Frankfurt, London, Amsterdam, Paris) but at lower cost than at present, as DC operators could potentially pass on savings made by water-based cooling system efficiencies.
If demand for AI increases as it has over the last few years, reducing the environmental burden it places on natural resources will be essential to ensure a long-term future for the technology industry and its users.
(Image source: “river” by cluczkow is licensed under CC BY 2.0.)
