Wood Mackenzie Says Carbon Capture Is Fastest Decarbonization Path as Global Data Centre Capacity Demand Reaches 200 GW by 2030

Wood Mackenzie has stated that carbon capture technology offers the most practical near-term pathway for decarbonizing the rapidly growing power demand from data centres, as natural gas continues to dominate global data centre energy expansion.

According to the company’s latest report series, “Decarbonizing data centres amid the gold rush for power: viable pathways,” all three major global combined-cycle gas turbine manufacturers are currently facing order backlogs while simultaneously expanding production capacity to meet rising demand from hyperscale data centre developments.

The report noted that global data centre electricity demand reached approximately 450 TWh in 2025, generating around 0.2 billion tonnes of CO2 emissions annually, equivalent to more than 0.5% of global emissions. While emissions from data centres remain lower than sectors such as steel, chemicals, and cement, the sector is experiencing significantly faster emissions growth driven by the expansion of artificial intelligence infrastructure and cloud computing facilities.

In the United States, data centres currently operate at emissions intensity levels around 48% higher than the national grid average, increasing pressure on hyperscalers and colocation developers to identify commercially viable decarbonization strategies.

According to Peter Findlay, the central issue is no longer whether gas-fired generation will power data centres, but whether those gas assets will be decarbonized.

The analysis found that integrating carbon capture technology with combined-cycle gas plants in the US could increase electricity generation costs by between USD 15 and USD 45 per MWh after federal 45Q tax incentives, bringing total power costs to around USD 115 per MWh. The report stated that carbon capture systems are capable of removing between 92% and 98% of flue gas emissions and can be deployed within three to four years or retrofitted to existing gas plants within three to five years.

Wood Mackenzie said the relatively short deployment timelines make carbon capture more commercially viable in the near term compared to alternatives such as new nuclear power projects, which typically require more than a decade to develop.

The report also examined emerging decarbonization technologies expected to become more competitive after 2030. Enhanced geothermal systems were identified as a potentially cost-effective long-term option, with projected generation costs as low as USD 61 per MWh by 2030–2035. However, the technology currently faces scalability limitations, with only around 1.5 GW in the development pipeline globally.

Nuclear power restarts were highlighted as another low-carbon option, although the report noted that only 11.5 GW of retired nuclear capacity is available for recommissioning in the United States. Small modular reactor projects were also identified as facing extended development timelines and cost uncertainties.

The analysis further stated that long-duration energy storage technologies remain significantly more expensive than competing solutions, with estimated costs ranging from USD 100 to USD 300 per MWh at scale.

While renewable energy and battery storage are expected to continue contributing to overall grid decarbonization, Wood Mackenzie noted that solar and wind generation alone may not be sufficient to meet the constant and high-reliability power requirements of hyperscale data centres without substantial gas-fired backup capacity.

The report concluded that hyperscale technology companies are increasingly balancing the urgent need for reliable power supply to support AI-driven growth with long-term sustainability commitments and decarbonization goals.