Subscribe to download Archive

Microsoft mulling advanced nuclear strategy

Microsoft is looking to power its datacentres with small nuclear reactors.
Microsoft is looking to power its datacentres with small nuclear reactors.

Tech giant Microsoft is getting more heavily into small-scale advanced nuclear power for its data centres. Founder Bill Gates is well known for having funded advanced nuclear research and start-ups, but Microsoft’s own interest till now had not been so clear.

Microsoft is hiring for a principal programme manager for nuclear technology to "be responsible for maturing and implementing a global small modular reactor (SMR) and microreactor energy strategy", reports Data Centre Dynamics.

Late in 2022, Microsoft had acquired Clean Energy Credits (CECs) from Ontario Power Generation (OPG) of Canada to power its data centres. The credits do include conventional nuclear power, but in the future could use the output of SMRs.

“This senior position is tasked with leading the technical assessment for the integration of SMR and microreactors to power the data centres that the Microsoft Cloud and AI reside on,” says the job listing. “They will maintain a clear and adaptable roadmap for the technology’s integration, diligently select and manage technology partners and solutions, and constantly evaluate the business implications of progress and implementation.” 

Data centres are energy-hungry and need steady power. Wind and solar are variable, and for a data centre would need to be backed up by ‘baseload’ power or batteries/long-term seasonal storage.

There is not enough green energy around to power the fast-growing number of data centres. Big tech companies typically have ESG and net-zero goals.

Microsoft is forward-looking with regard to nuclear power. In May, Microsoft bet on fusion power by signing a power purchase agreement (PPA ) with Helion Energy for electricity generated from Helion’s first fusion power plant.

The plant, to be sited in the US, is to come online by 2028 and will have an installed capacity of 50 MW after a one-year ramp-up period. However, sceptics says the technology is potentially years away from being proven.

Similarly, SMRs are unproven commercially. Their proponents say they will be able to be mass-produced, keeping costs down and allowing developers to stick to deadlines. Conventional large-scale nuclear is typically over-budget and late.