The carbon cost of AI Green goals hit hard by drive to develop technology

If you want evidence of Microsoft's progress towards its environmental "moonshot" goal, look closer to Earth: at a building site in west London.

- Dan Milmo, Alex Hern

The company's datacentre in Park Royal is part of its commitment to drive the expansion of artificial intelligence (AI), one that jars with its target of being carbon negative by 2030.

Microsoft says the centre will be run fully on renewable energy. But building datacentres and the servers they are filled with means the company's scope 3 emissions - such as CO₂ related to the materials in its buildings and the electricity people consume when using its products are more than 30% above their 2020 level. As a result, it is exceeding its overall emissions target by roughly the same level.

This week, Microsoft's co-founder, Bill Gates, claimed AI would help combat climate change because big tech is "seriously willing" to pay extra to use clean electricity sources in order "to say that they're using green energy".

But in the short term, powerhungry AI is problematic. Brad Smith, Microsoft's outspoken president, once called its carbon ambitions a "moonshot". In May, stretching that metaphor to breaking point, he admitted that because of its AI strategy, "the moon has moved". It plans to spend £2.5bn over the next three years on growing its AI datacentre infrastructure in the UK and this year has announced new projects round the world, including in the US, Japan, Spain and Germany.

Training and operating the AI models that underpin products such as OpenAI's ChatGPT and Google's Gemini uses a lot of electricity to power and cool the associated hardware, with additional carbon generated by making the related equipment.