31 Unattributed quote 31 AI and energy investment need to be aligned, building data center capacity in countries which are energy rich, providing access to compute capacity to countries who have the technology capabilities but not the infrastructure and generation capacity to keep up, while finding the right balance between compute capacity and energy for countries with limited foundations. The right partnerships will be needed between energy and technology leaders, with economic models developed to understand the economic benefits versus where the revenue is generated, which may not always be where the infrastructure sits. In the AI race, countries risk offering incentives and tax benefits which result in limited domestic economic benefit if sustainable jobs are not created, water scarcity impacts agriculture and there is a lack of access to affordable energy for all. On the other hand, the energy demand of today’s hyperscalers can be equivalent to the size of a city supporting 4m households. This can enable vast generation capacity to be built in one go near cities which lack reliable access to energy, built ready to meet future demand growth where otherwise the building at scale would have been uneconomic. This could unlock otherwise unfeasible hydroelectric, geothermal, nuclear or gas projects in places where the source of energy is abundant, unlocking the full potential of the country’s energy ecosystem. The benefits of data centers could be significant with the right planning, integrating water, heating and cooling for district and agricultural use. Building power generation capacity with co-located data centers built with redundancy could also provide low-cost non-guaranteed domestic power (which will inevitably be better than the current availability). We’ve cut permitting from years to weeks. Environmental Impact Statements that once took four years are now completed in under 28 days. When we accomplish this at scale, capital starts flowing — you can’t invest if you can’t get approval.