24 25 Over the last year, hyperscalers have transformed from projection to reality, building at scale and pace. Global installed capacity stands at 122 GW, 80% of which is concentrated around the US (53.7 GW), China (31.9 GW), and the EU (11.9 GW). The US has added 12 GW of new capacity in 2025, positioning the market to reach 98 GW by 2030. This is no longer incremental growth—it is structural hyperscaling. Hyperscale campuses are equivalent to adding entire cities of power demand exclusively for compute, driving demand for firm power solutions, advanced cooling, and new grid architectures. This concentration of capacity is forcing utilities and governments to rethink grid planning, permitting, and energy procurement strategies. The scale is unprecedented. Meta’s Hyperion campus will scale to 5 GW by 2030, supporting AI workloads at unprecedented density, while OpenAI’s Stargate project is targeting 26 GW across five US sites—comparable to the entire power demand of Greece. Building Infrastructure at Scale and Pace 5 Planning the next wave for greater integrated benefit This concentration of demand has led to a rethink in energy supply, shifting from incremental renewable PPAs to firm baseload solutions. This has led to a renewed dash for gas not seen since the US switched from coal to gas in the early 2000’s, and a nuclear renaissance as data center operators look for clean dispatchable energy at utility scale. The ability of the grid infrastructure to keep up with this unprecedented demand growth has led to data center operators to look at ‘behind the meter’ or ‘islanded’ supply solutions. US retail power prices have increased on average by 23% between 2020-2024, leading to a weakening of the social licence with households. While data centers’ share of power demand in the US, Europe and China is forecast to rise from 2% today to 10.5% in 2050, it will represent less than 25% of total power demand growth within this region. The remaining demand growth will come from other sectors such as industry, buildings and transportation which will be more dispersed and incremental in the nature of its growth. See Fig. 10 Figure 10 Figure 11 Data center electricity demand is expected to more than double by 2030, adding over 500TWh- more than the consumption of Japan. However, beyond this, the outlook is far less certain around the demand for digital services, the shift from training to inferencing and the rapid improvement in efficiency of compute and data centers themselves. While data center operators have brought a new wave of investment to base load generation in the US, Europe and China, 50% of power demand growth to 2050 will be outside of these regions. This short-term burst of demand will have ripple effects on future power supply strategies in other regions and is already showing signs of significantly accelerating the viability of new-generation technologies such as SMR, fusion and geothermal. Siting of the next wave of data centers will need to learn from the lessons of the last 12 months to ensure that the integrated benefits of AI fully serve the communities in which they are placed, unlocking low-cost energy, creating jobs, and raising economic wealth. Forecasting the amount of compute the world needs remains a problem that even AI cannot solve today. See Fig. 11. While 80% of today’s data center power consumption is being used to train large language models, it is anticipated that this will drop significantly, with 80% of the demand shifting to inferencing, as billions of people and corporations leverage these productivity tools. Sub-Saharan Africa alone will add 600m new digital citizens over the next five years. Dealing with data center uncertainty Data sovereignty, accuracy and latency will all influence the size and location of future data centers. Data partnerships could enable greater sharing of data center capacity amongst smaller nations, while latency will create natural economic barriers (every 1,000km adds 5-7ms, limiting service geography for some applications). While many tasks will require high levels of accuracy, new, low-energy models such as the UAE’s K2 Think are changing the amount of energy needed for simpler tasks and queries. Marginal gains can be achieved by increasing AI accuracy from 88% to 96%, but at the cost of 28x more energy and 40x longer response times. Semiconductor chips are becoming increasingly more efficient. AMD and Nvidia chips deliver 2x performance per watt while Google’s Ironwood AI chip has improved efficiency by 50%. At the same time, progress on quantum computing promises to revolutionize the energy needed in the computation of complex tasks. The energy consumption of data centers will depend on the power usage efficiency (PUE) which measures how much energy is needed per GW of compute. The global average today is 1.56–1.58, but the best hyperscale sites are achieving 1.1–1.2, with Google and AWS deploying liquid cooling and immersion systems across new AI campuses, while Meta’s Prometheus and Hyperion sites use direct-to-chip cooling for efficiency. These innovations are helping to also improve the water usage efficiency (WUE) from the industry average of 1.8 l/kWh to advanced designs ≤0.2 l/kWh. Existing hyperscaler energy demand is expected to rise as rack density increases. A decade ago, typical racks consumed 6–8 kW of power. Today, AI racks draw 130– 250 kW, and projections suggest some could reach 1 MW per rack- enough to power a small neighbourhood. Hyperscalers like Meta and Microsoft are designing campuses to handle these ultra-dense GPU clusters, which will contribute to further improvements in PUE but require higher localized energy demand, adding to the demand for base load generation. Unattributed quote 8% 15% 3000 Low-High Range 2500 2000 1500 500 1000 0 2050 10% Global IT Computing Capacity (GW) CAGR 2024-2050 Base-case 2050 2045 2035 2030 2024 2040 DC % of Total Power Generation 2024 2030 2040 2050 +73% DC Transport Buildings Industry 1.8% 4.1% 7.6% 10.5% Gas % of Total Power Generation 13.9% 13.6% 11.8% 8.6% Concentrated Power Demand Dispersed, Incremental Power Demand 2025-2030 2030-2040 2040-2050 Others Data Centers 2.6 16% 82% 22% 25% 78% 75% 0.6 3.1 3.8 3.9 1.1 1.3 4.9 5.3 Power Demand US, Europe and China, ‘000 TWh Employment in LICs stands to gain and lose relatively less Drivers of development differences between LICs and HICs* Current USD (tm) Employment shares by AI exposure and complementarity High exposure, low complementarity Low-income countries High-income countries High exposure, high complementarity Low exposure Power Demand Growth US, Europe and China, ‘000 TWh 200 180 160 140 120 100 80 60 40 20 0 $75tm 16 53 22 20 13 62 186 LICs* (today) *LICs = Low-income countries . HICs = High-income countries. % of labor force LICs* (at HICs levels) Energy Finance Education Health Governance 0 20406080 100 18 33 27 40 8 74 DC % of Total Power Generation 2024 2030 2040 2050 +73% DC Transport Buildings Industry 1.8% 4.1% 7.6% 10.5% Gas % of Total Power Generation 13.9% 13.6% 11.8% 8.6% Concentrated Power Demand Dispersed, Incremental Power Demand 2025-2030 2030-2040 2040-2050 Others Data Centers 2.6 16% 82% 22% 25% 78% 75% 0.6 3.1 3.8 3.9 1.1 1.3 4.9 5.3 Power Demand US, Europe and China, ‘000 TWh Employment in LICs stands to gain and lose relatively less Drivers of development differences between LICs and HICs* Current USD (tm) Employment shares by AI exposure and complementarity High exposure, low complementarity Low-income countries High-income countries High exposure, high complementarity Low exposure Power Demand Growth US, Europe and China, ‘000 TWh 200 180 160 140 120 100 80 60 40 20 0 $75tm 16 53 22 20 13 62 186 LICs* (today) *LICs = Low-income countries . HICs = High-income countries. % of labor force LICs* (at HICs levels) Energy Finance Education Health Governance 0 20406080 100 18 33 27 40 8 74 We need all the technology we can get — more natural gas, more nuclear, more solar, more wind — and we need it with stability. The one thing we can’t afford is to go in circles every election cycle. The price of electricity is now a political issue.