AI hardware lifespan threatens profits and the environment

Investment in artificial intelligence has surged, driving rapid growth in chips and data‑center capacity. Analysts now warn that the usable life of the hardware that powers AI models may be considerably shorter than expected, which could cut profit margins for firms and pressure their market valuations.

At the same time, the operation of AI systems requires vast amounts of electricity, continuous cooling that consumes millions of litres of water, and the production of devices that depend on minerals such as lithium, cobalt, nickel, copper and rare‑earth elements. The extraction and processing of these resources impact ecosystems and raise serious sustainability questions for the AI sector.