AI opens a new turning point in European electricity consumption, and giant utilities companies are welcoming the wave of "valuation expansion".

The Wall Street financial giant Goldman Sachs Group, Inc. released a research report stating that European undervalued utility giants focusing on electricity supply and water resource supply networks will greatly benefit from the unprecedented global AI boom and the continuing trend of electrification for many years. Goldman Sachs Group, Inc. stated that over the past fifteen years, Europe's electricity demand has been severely impacted by external shocks (such as the global financial crisis, COVID-19, and energy crises after the Russia-Ukraine conflict) and slower-than-expected electrification processes, resulting in a cumulative decrease of about 10% in electricity consumption since 2008. However, Goldman Sachs Group, Inc. predicts that this trend is about to be completely reversed: the rapid expansion of data centers and the gradual acceleration of electrification processes may push European electricity demand to increase by at least about 40%-50% in the next ten years. In the research report, Goldman Sachs Group, Inc. mentioned that at a technical exchange meeting in December last year, the British electricity giant Centrica confirmed the potential to achieve 1.6 billion in EBITDA by the fiscal year 2028 (this data is the median in the range, with performance guidance from 1.3 billion to 1.9 billion), of which 400 million will come from Centrica's 4 billion investment plan. Currently, around 50% of this 4 billion has not been committed and there are various alternative options. Significant investment projects emphasized by Centrica include new nuclear power, hydrogen energy, and carbon capture. Goldman Sachs Group, Inc. stated that in terms of nuclear power, details of Centrica's investment may be disclosed in an expenditure review on June 11, 2025. Although the details of the regulatory mechanism will be crucial, Goldman Sachs Group, Inc.'s analysis team provided an explanatory analysis to help assess the scale of this opportunity for Centrica. For example, a 2 billion equity investment, if it meets the midpoint of Centrica's target return threshold (7-10%+ internal rate of return IRR), calculated at a 6.5% cost of equity, could potentially bring about a 6% upside in the current stock price and increase earnings per share (EPS) by about 3% by 2028 under unchanged conditions. Goldman Sachs Group, Inc.'s research team conducted a detailed analysis of various scenarios, continuing to view the electricity giant Centrica as a strong cash generation story, with its business transforming towards green investments, data centers, and electrification. Reinvesting through its financial space on the balance sheet at a target return rate (7-10%+ internal rate of return IRR) helps improve the business's valuation multiples. Goldman Sachs Group, Inc. maintains a "buy" rating for Centrica, with a revised target price of 190 pence/share, implying a potential upside of about 20%. It is understood that Centrica's core business covers retail natural gas and electricity, upstream oil and gas, flexible peak power stations and energy storage, and holds a 20% stake in British nuclear power stations; evaluating new green assets such as nuclear, hydrogen, carbon capture. The rapid expansion of AI data centers and electrification trends is expected to boost electricity demand, driving significant growth in wholesale electricity prices and electricity capacity markets, providing significant room for Centrica's EBITDA and EPS to thicken. The end of AI is electricity! As ChatGPT, Claude, DeepSeek, and other artificial intelligence applications sweep across the globe, the energy demand of global mega-scale AI data centers is so huge that some utility stocks that have long been overlooked by the market have entered the sights of top Wall Street investment institutions this year. The accelerating AI power demand closely associated with artificial intelligence training/inference is expected to continue to surge in the coming years as data centers, which are already "power monsters," expect a substantial increase in their need for electricity resources. AI data centers can be said to be the most critical large infrastructure construction project in the AI era, crucial for the efficient operation of generative AI applications such as ChatGPT and the iterative updates of large AI models like the GPT series. Behind the exponentially expanding high-energy AI data centers driven by the fierce demand for AI infrastructure such as AI chips, electricity supply is the core foundation, which is the origin of the market view that "the end of AI is electricity". A forecast report from the International Energy Agency (IEA) shows that by 2030, global data centers' electricity demand will more than double, reaching around 945 terawatt-hours (TWh), slightly higher than Japan's current total electricity consumption, with artificial intelligence applications being the most important DRIVING force behind this growth; it is estimated that by 2030, the overall electricity demand of data centers focusing on artificial intelligence will increase by more than four times. According to forecasts from the Boston Consulting Group, the share of data centers in electricity consumption in the United States is expected to double, increasing from 126 TWh in 2022 to 390 TWh in 2030. TWh is used to describe the largest levels of electricity consumption and is usually used for national energy statistics, as well as for planning and evaluation of large-scale energy projects. Large industrial facilities, such as super steel mills, may consume less than 10 TWh of electricity in a year. PJM Interconnection LLC, which manages the power systems of 13 states from Washington D.C. to Illinois, recently predicted that due to the booming trend in the field of artificial intelligence, the summer peak electricity demand is expected to increase by nearly 58 gigawatts (approximately 38% growth) from last year's historic high by 2035, reaching around 210 gigawatts. This expected increase in electricity demand is more than twice the highest demand in the history of the New England power grid. With a sharp increase in the number of new large data centers used to run AI large models and a significant expansion in their scale, it is expected that the electricity consumption of PJM and the United States as a whole will grow at an unprecedented speed in the next decade.There is a significant increase in power consumption. Predictive data shows that some planned electricity infrastructure projects may consume enough energy to supply the entire city's electricity needs.Despite the Biden administration and the current President Donald Trump led U.S. government both claiming the need to increase efforts to deploy artificial intelligence infrastructure to maintain national security and economic interests, Wall Street analysts widely predict that the significant increase in electricity demand is likely to put immense pressure on the aging infrastructure and supply scale of American Electric Power Company, Inc. year by year. This is also why Wall Street funds have been pouring into power stocks such as Vistra Energy (VST.US) since 2024. The U.S.-based power giants Vistra Corp. undoubtedly had the best performing stock in the 2024 S&P 500 Index, with an increase of up to 256%, surpassing even NVIDIA Corporation's 170% increase in 2024. Vistra's performance this year has also surged by 30%, almost reaching a historical high, significantly outperforming the S&P 500 Index's 1% increase. Vistra and its power counterparts Constellation Energy Corp (CEG.US) have shown remarkable growth in the U.S. utility sector, as they are positioned as independent power producers, meaning they sell electricity at market prices, unlike some regulated utility companies in the U.S. It is reported that Constellation Energy announced on Tuesday that it had reached a 20-year power purchase agreement with Meta (META.US), the parent company of Facebook, to purchase power from the Clinton Nuclear Power Station starting in mid-2027. This agreement once again confirms the huge power demand of data centers owned by major tech giants. In addition to power giants, utility giants focused on water resources are also favored by Goldman Sachs Group, Inc. Pennon Group released its 2025 financial performance on the morning of June 3rd local time, with core performance data including basic EBITDA of 335.6 million, basic EBIT of 148.5 million, and adjusted earnings per share (EPS) of -10.3 pence. Goldman Sachs Group, Inc. stated that in the 2025 financial year, Pennon achieved a net result delivery incentive (ODI) penalty of approximately 23 million (involving South West Water and Bristol), with a cumulative regulatory asset return rate (RoRE) of 6.0% for South West Water in the 2025 financial year, reflecting a total capital expenditure (totex) contribution of -1.6% and a result delivery incentive (ODI) contribution of -0.6%; the company's net debt at the end of the 2025 financial year was 4.1 billion, which is approximately 85 million higher than analysts' expectations compiled by Bloomberg. Overall, Goldman Sachs Group, Inc. believes that Pennon Group's financial performance in the 2025 financial year meets Wall Street expectations, and the new performance outlook provided is broadly in line with the 2026 financial year consensus expectations on Wall Street. The net debt was slightly below market expectations, with an excess of 87 million compared to consensus expectations (only about 4% of the market value). Therefore, Goldman Sachs Group, Inc. maintains a "neutral" rating on Pennon and continues to set a target price of 554 pence, indicating a potential upside of nearly 20% in the future. Pennon Group, the leader in water supply in the U.K., focuses on water supply and sewage treatment networks, indirectly benefiting from the enormous cooling water, humidification, and preparation of high-purity water needed for data center server farms through wastewater recycling, industrial cooling water supply, and waste heat utilization projects, continuously reaping strong cash flows through the regulatory return model. High-performance AI server clusters and giant cloud computing platforms for AI training/inference are rapidly expanding, causing data centers to simultaneously surge in both the "electricity-water" resource chains. Based on the latest calculated data disclosed by enterprises, a single large-scale hyperscale data center can consume 3 to 5 million gallons of drinking water daily, used for liquid cooling systems, humidification, certain living facilities on the premises, and even recycling water systems, equivalent to the daily demand of a city of 30,000 to 50,000 people; when converted based on power capacity, a 1 MW data center requires 25.5 million liters per year only for cooling (approximately 67,000 gallons/day), with an industry average Water Usage Efficiency (WUE) of approximately 1.8 liters/kilowatt-hour. Therefore, the demand for water resources by large data centers driving AI training/inference systems is "vast and concentrated", mainly reflected in evaporative cooling, breakthrough liquid cooling systems, humidification, and the preparation of high-purity water needed for data centers; the unprecedented surge of AI computing power further amplifies this demand, prompting utility operators to accelerate the deployment of new water resource utilization technologies such as liquid cooling systems and reclaimed water utilization.