Artificial intelligence is changing the world’s electricity needs. Data centres need power around the clock, and the rapid build-out is putting pressure on grids that are already strained. When the session CCS and Decarbonising AI Data Centres opened in the IETA pavilion at the climate summit in Antalya on 13 November, moderator Noora Al Amer of the Global CCS Institute pointed to estimates that data centres could account for 7 per cent of electricity demand.
Mona Mølnvik, Research Director at SINTEF Energy Research, took part in a panel discussion together with Ilaria Parrella, Global Sales Director for Climate Technology Solutions at Baker Hughes.
Gas power with CCS is a viable solution
“For data centres, gas power with CCS is a viable solution,” said Mølnvik. Gas-fired power plants being built for data centres now should include a CCS solution. CCS has been proven at full scale today, she stressed.
The challenge is that every CCS-plant so far has been tailor-made. Data centres need a standardised solution that can be built identically for many facilities.
“The flue gas from gas turbines contains only around 4 per cent CO₂, which is lower than industrial sources. SINTEF researches compact heat recovery units that integrate CO₂ capture with the heat of a gas power plant, such as a combined-cycle plant, to raise efficiency and make the solution more compact. Once the learning curve kicks in, because more identical plants are built, the cost falls,” she said.
Mølnvik also pointed to Brevik, where retrofitting CO₂ capture on a cement plant more than a hundred years old was complicated – but succeeded nonetheless.
“The opportunity to streamline CCS for data centres from the outset is exciting,” she said.
Parrella explained that Baker Hughes, which works on CCUS, hydrogen and geothermal energy, sees carbon capture as a way to deliver low-carbon power at the pace data centres demand, and that the company is involved in a gigawatt-scale project where CCUS is planned as an integrated solution from the start.
Think chains and hubs
“Some full-scale projects are ready on paper, but stall when the final investment decision has to be made,” said Mølnvik.
The answer is to plan the whole chain: capture, transport and permanent storage – and then combine CO₂ from the data centre’s power plant with CO₂ from other nearby industry in shared hubs. Several sources sharing transport and storage lowers the total cost.
Norway must also discuss alternative solutions
Norway, which is attractive for data centres because of its cool climate and clean power, does not have enough electricity for everyone. She noted that the Norwegian government recently signaled its intention to introduce a licensing scheme for existing and future data centres.
“So in Norway, too, we need to discuss alternative solutions,” she said.
Research made full scale possible
On collaboration, Mølnvik pointed to the research centre FME gigaCCS, with around 40 partners, many PhD candidates and international university collaboration. The key is to bring industry into the research early.
“Without the research, the full-scale projects would never have happened,” she said. Her message to policymakers: net zero is impossible without CCS, and the whole world has to be on board.
“That means we need agreements on transporting CO₂ across borders, and policies that make it expensive to emit CO₂. One example is the EU’s carbon border tax (CBAM), which could drive CCS forward outside Europe,” she said. For example, Petrobras is already working with ArcelorMittal to decarbonise steel production in Brazil for export to Europe.
Responding to a critical question from the audience, Mølnvik added that data centres should not be built uncritically – the debate about how many Norway should have, and what they should be used for, is legitimate. And asked whether CCS actually works, she pointed to open data: storage data from Sleipner, where CO₂ has been injected since 1996, is freely available through the SINTEF-led portal CO₂ DataShare.
Monitoring and traceability of CO₂ remain demanding, which is why SINTEF is building reference data and researching fiscal metering of CO₂ – work that is well under way.
Mølnvik is co-author of the CCS chapter in Fra ambisjon til gjennomføring (From Ambition to Implementation), NTNU and SINTEF’s expert advice for COP31.