8 9 1. The Energy Outlook Energy companies are struggling to keep up with the current surge in demand and as such the world must focus on doing more with less energy. Electrification is key to a lower energy intensity world, where EVs and heat pumps are far more efficient than their predecessors but need a build out of new infrastructure to support them. 80% of today’s data centers are concentrated within the US, Europe and China, relying on 26.4% gas to generate the power they require, using location-based accounting. With rising gas prices this has resulted in higher consumer power prices, while shortages and delays in the gas turbine supply chain are creating a new choke point. Energy and AI are the twin engines that have the potential to turbo-charge socioeconomic growth, but energy-intensive data centers are competing for energy alongside multiple demand drivers in a growing and increasingly complex world. Gathering at the third ENACT Majlis, energy, technology, investment and government leaders met in Abu Dhabi to advance, together, the integrated solutions needed to ignite the twin engines of energy and AI to accelerate global growth. This Majlis emphasized a pragmatic, “less rhetoric, more results” approach, an “and-and” energy strategy across molecules and electrons, and urgency on permitting, supply-chain diversification and talent to keep pace with AI demand. Executive Summary Electrification Scale Data centers Concentrating Power Demand • Gas filling the gap • New technologies ( nuclear fission) Electrification Incremental Energy Efficiency Dispersed, Incremental Power Demand Renewables Capital Flow Wealth Creation Policy (Governance) People (Talent) Energy As a new, energy-intense sector, data centers can be the solution for bringing new capital to energy markets and concentrated demand to enable investment in highly capital-intense projects such as gas turbines, nuclear fission, geothermal and even fusion. Data centers will need to be sited around regional archetypes based on access to abundant low-cost energy, talent, and resources able not just to consume but also drive build-out, integrated value such as water, cooling and heating (into residential, agriculture etc.) and availability of capital. Participants added three siting imperatives: co-locate compute with generation to bypass transmission bottlenecks; pair gas with renewables and batteries to deliver cleaner-firm profiles; and engage communities early on water and price impacts to sustain a social license to build. However, we must be clear that efficiency gains alone will not overcome grid constraints. Majlis participants highlighted immediate, scalable efficiency wins: governments enabling private performance contracts can yield 20–30% savings in electricity and water at system level, freeing capacity for AI loads before new plants are needed. We still rely on molecules to create the electrons that AI needs. This is not an energy transition, but an energy addition — more gas, more renewables, more storage to power data centers reliably. Unattributed quote Over the past 12 months, the US alone has added 12GW of data center capacity, which has added to the surge in electrification around the world and which, today, still relies on 57% molecules to produce that power. Participants acknowledged that today, we are less in an energy transition than in an energy addition, needing more gas, more renewables and more batteries as the only way to provide firm power to data centers. Massive uncertainty in energy demand depends on the pace of AI growth, development of the Global South, and the speed of the energy transformation, with power supplied from molecules expected to fall to 21% by 2050. Building on this, participants stressed that “power is knowledge”: without abundant, affordable electricity, AI’s promise stalls. Policy stability is as critical as geology, and capital will follow low-cost power. In addition, the industry is facing price pressures which could impact capital budgets. Lower oil prices may impact production, including in the US, where lower associated gas would impact availability and push up the price of gas for data centers, and would decelerate the progress in shifting to alternative energy sources such as EVs and SAF. Lower investment could lead to future price shocks. Participants also warned that public tolerance for higher electricity bills is shrinking, making affordability a political risk for AI build-outs unless new generation and grid upgrades are co-delivered with siting. For the first time in human history, a kilowatt of electricity can be turned directly into intelligence. But we cannot achieve the dream of AI without more power. Power is knowledge now. Unattributed quote 2. Doing More with Less 3. Data Center Siting to maximize integrated value for data centers and consumers in a more energy efficient world dealing with uncertainty and price pressures