Comments: Grabbing the Brass Ring To Power the Demand for Data Centers and Generative AI

Abstract

“There’s no way to get there without a breakthrough.” These were the words of OpenAI’s CEO Sam Altman at a sideline meeting with Bloomberg at the World Economic Forum (WEF) in Davos, Switzerland, in January, referring to the energy required to power generative AI, data centers, cloud computing, and to support required equipment and infrastructure. As industries swiftly transition into a fresh era of digital revolution, spearheaded by the fast adoption and advancement of generative AI technology, the demand for energy to power data centers and required infrastructure skyrockets. Research firm IDC estimated that global data center energy consumption reached 382 TWh in 2022, with a forecast surge to 803 TWh by 2027. To contextualize, 382 TWh is roughly equivalent to France's annual electricity consumption, while 803 TWh is comparable to Russia's usage (International Energy Agency, 2019 data). Currently, data centers consume approximately 1 to 2% of global electricity, is expected to rise to 8% by 2030, according to WEF. Struggling to keep pace are the sources of energy required to fuel this transformation, and the transmission and distribution constraints of power grids. In 2010, IDC observed a 50% gigaleap in global data production to 1.2 zettabytes (1.2 trillion GB), with projections of annual data production reaching 35 zettabytes by 2020, a milestone reached in 2018 instead. By 2020, 59 zettabytes of data were generated. IDC anticipates newly created data to soar to 175 zettabytes by 2025, notching a 146-fold increase over 15 years. The voracious energy consumption of data centers is sounding alarms about the need for additional sustainable and efficient power generation sources. SPE-218905, presented last month at the 2024 SPE Western Regional Meeting, dove into the oil and gas industry’s role in assuring “digital decarbonization.” Noting the expanded use of AI, data centers, and cloud computing in our industry, the authors wrote, “The need for energy transition has triggered an unprecedented momentum for developing renewable energy sources. With its historical and current contributions to the energy and products market, all indications are that the petroleum Industry needs to continue its efforts in the production of hydrocarbons while also actively exploring ways to substantially reduce GHG emissions such as methane and carbon dioxide.” Major companies are intensifying digital transformation efforts, reshaping petroleum industry processes. This includes reservoir characterization, drilling design, and economic calculations, which rely on high-performance computers—increasing energy consumption and emissions. Data analytics and AI aid in demand forecasting, inventory management, and predictive maintenance, reducing disruptions and environmental damage, thereby cutting carbon emissions. Automation, drones, robotics, and IoT sensors enhance operations, safety, and efficiency while minimizing environmental impact. Data reign supreme as the "ultimate authority" across the entirety of our industry's operations, guiding the evolution of our practices to foster cleaner operations, prevent downtime, and gain efficiencies. The authors added this is crucial for remaining competitive and improving the industry's reputation, particularly amidst stringent environmental regulations in the US. They recommend actions to advance the sustainable implementation of digital technologies while also conserving power consumption and minimizing emissions. Among the suggestions are power management systems to track and optimize energy usage, the use of renewable and low-carbon energy sources, prioritizing partnerships with cloud providers committed to sustainability, and improving hardware sustainability by using durable materials and energy-efficient designs throughout its life cycle. Leading the development of breakthroughs are operators exploring alternative energy sources and their implementation. Diamondback Energy, the largest independent producer in the Permian Basin, is working with Oklo to develop fast fission nuclear reactors. SPE 215407, presented at the 2023 SPE/IATMI Asia Pacific Oil and Gas Conference, described using low-cost green power for unconventional oil recovery and sharing energy-storage facilities with data centers. A hybrid solution integrates a traditional grid with mobile microgrids between shale-oil fields and data centers. Developing alternative power supplies with wide-scale reliability, dependability, and minimization or elimination of GHG emissions within a feasible capex/opex scenario is the brass ring of sustainability and energy security—and data are helping us get there. Data Centers Around the World: A Quick Look by Brian Daigle, US International Trade Commission. SPE 218905 - Digital Decarbonization in the Oil and Gas Industry by I. Ershagi, University of Southern California; Donald L. Paul and Andrei Popa, Chevron Technology Center. https://doi.org/10.2118/218905-MS SPE 215407 - Unconventional Fields Recovery Enabled by Large-Scale Green Power Supply Based on Multi Microgrids and Energy-Storage Sharing With National Data Centers by Z. Tong, X. Wang, and D. Weng et al., PetroChina. https://doi.org/10.2118/215407-MS Data Science Has Exploded Across the Industry Over the Past 75 Years by Adam Wilson, JPT.