The increasing use of AI could cause data center energy consumption to triple by 2030.
“In the last two decades, no form of technology has fueled the need for accelerated infrastructure development in Europe like AI,” analysts write in McKinsey’s report. Today, data centers in Europe consume about 62 Terawatts per hour (TWh), but that could rise to 150 TWh. That would be five percent of the entire European energy chain.
Big challenge
To handle that increase would require an additional 25 gigawatts of green energy. That conflicts with the goal of becoming carbon neutral by 2040. One possible solution is to place data centers close to clean energy sources. This is true for AI training models that do not need to be close to large cities. The more AI applications require computing power and low latency, the closer they should be to urban areas.
McKinsey suggests carbon capture and small-scale nuclear reactors on-site as a long-term solution. In addition, diesel generators, fuel cells and batteries can also support the energy demand of data centers. Surprisingly, McKinsey does not think energy will rise that much because the EU saves enough on energy. Milder winters and de-industrialization due to high energy prices also help stagnate consumption.
Research firm Gartner predicted that the use of new and different processors will significantly reduce energy consumption. Unfortunately, that decrease is complex and very expensive, so that investment must be well worth it.
read also