At the end of artificial intelligence is really electricity! Microsoft Corporation (MSFT.US) made a bold bet of 70 billion dollars, heralding the "self-powering age".

US technology giant Microsoft Corporation (MSFT.US) is currently in exclusive negotiations with US oil and gas giant Chevron Corporation (CVX.US) and renowned investment fund Engine No. 1 to reach a long-term agreement to provide strong support for a large energy complex in western Texas and to power a super-sized AI data center park led by Microsoft Corporation. According to reports citing insiders, the planned mega natural gas power plant in western Texas is expected to cost around $7 billion, with an initial power capacity of 2500 megawatts, making it one of the largest projects of its kind in the US. As negotiations are still ongoing, the sources have requested anonymity. In a jointly released statement, the companies said, "Chevron Corporation, Microsoft Corporation, and Engine No. 1 have signed an exclusive agreement regarding a proposed power system and power purchase arrangements. The commercial terms are not finalized yet, and there is no final agreement at this time." Chevron Corporation and Engine No. 1 previously disclosed some details of their planned large power plant but did not reveal the final user. If an agreement is reached with Microsoft Corporation, it would potentially secure a large long-term customer for the power plant and help support its financing and funding. The project could potentially start operation as early as 2030, pending approval from US tax and environmental authorities, as well as agreement on commercial terms between the three companies. It was reported that the Trump administration in February publicly urged large tech companies to "be responsible for meeting their own power needs." Subsequently, the White House pushed for commitments from companies like Microsoft Corporation, Alphabet Inc. Class C, Amazon.com, Inc., and Meta (Facebook's parent company) to independently bear the related costs of energy infrastructure and to increase new power supply. The government's policy direction of having tech giants generate their own power is actively promoting the investment theme that "the end of AI is electricity." The US artificial intelligence technology industry seems to be entering a new phase of "actively grabbing power" and "self-supplying power." For global tech giants like Microsoft Corporation, Alphabet Inc. Class C, and Amazon.com, Inc., energy resources are increasingly becoming core competitive barriers in the AI era. On one hand, the US power grid is facing challenges such as grid congestion, transmission bottlenecks, and shortages of large unit equipment. PJM, one of the largest grid operators in the US, has stated that the demand for data centers is growing faster than new supply, and a power shortage could potentially occur as early as next year. On the other hand, tech companies closely associated with artificial intelligence technology have progressed from simply signing power purchase agreements (PPAs) to directly tying to underlying power assets. Microsoft Corporation, through its agreement with Constellation, is driving the restart of the Three Mile Island nuclear power plant to provide power for its large data centers. Meta has agreed to pay all service costs for its super-sized data centers in Louisiana and is striving to push for over 5,000 megawatts of clean power resources in Texas, leading the construction of its largest data center globally in Richland Parish. The recent actions taken by these tech giants seem to indicate that whoever can quickly secure base load power sources, underlying large-scale transmission channels, and unit resources will be more qualified to expand the next round of AI computing power. The concept of "building a power system oneself" in reality usually does not mean completely disconnecting from the grid or creating a standalone power system but rather tech companies using their own funds to secure power sources and bear the costs of grid connection and supporting infrastructure, and when necessary, construct exclusive or semi-exclusive power supply projects. A $7 billion investment igniting the self-supplying power era! Microsoft Corporation joins forces with Chevron Corporation in a bold move in the Texas energy sector Microsoft Corporation, as a longtime supporter of OpenAI, the developer of ChatGPT, is persisting in building data centers despite recording its worst quarterly performance since the 2008 financial crisis. This move is aimed at continuously competing with Alphabet Inc. Class C's parent company Alphabet Inc. and Amazon.com, Inc. for dominance in the cloud computing sector and absolute leadership in cloud-based AI training/inference platforms. Obtaining reliable base load power has become a crucial challenge, and the collaboration between Chevron Corporation and Engine No. 1 is expected to address this challenge with their extensive natural gas production and large turbine contract projects in western Texas. The next generation of large data centers often choose to be built in areas far from major population centers but closer to fuel sources due to their enormous power requirements. The project location selected by Chevron Corporation and Engine No. 1 near Pecos, in close proximity to the Texas-New Mexico border, is situated in the heart of the Permian Basin, the largest oil and gas producing region in the US. The natural gas produced in the Permian Basin, a byproduct of oil production, is so abundant that it often exceeds pipeline transportation capacity. As a result, some natural gas is burned directly on-site because it cannot be transported to demand locations, making the region an ideal location for building power plants. Landbridge Co. LLC, a major land resource owner in the area, stated that in the past two years alone, there have been at least nine large data center projects proposed by tech companies in northern and western Texas. A key strength of Chevron Corporation's entry into the AI power supply sector lies in its long-term partnership with Engine No. 1. The latter is a shareholder activist hedge fund that initiated successful aggressive investment actions against its rival Exxon Mobil Corporation in 2021. This partnership has secured seven large gas turbine orders from GE Vernova, with a high demand for such equipment resulting in new customers facing waiting periods of several years. Chevron Corporation stated that the large power project in Pecos is expected to be operational by 2027 and will gradually increase to a power capacity of 2500 megawatts within three years. According to data from the U.S. Energy Corp. department, this is equivalent to the power generation capacity of more than two typical nuclear reactors. If successful, the large power plant could potentially be expanded to a maximum capacity of 5000 megawatts in the future. Operating under the entity name "Energy Forge One LLC," Chevron Corporation has applied for several tax incentives from local authorities in western Texas. Its air emissions permit application submitted to the Texas Commission on Environmental Quality was deemed administratively complete in October. The end of AI is electricity! A "super bull market" for electric power stocks is on the horizon. As the Trump administration pushes US tech giants into the era of "self-supplying power," Microsoft Corporation, Alphabet Inc. Class C, and Meta are all recently engaging in long-term power purchase agreements, co-locating gas turbine units, restarting nuclear power plants, supporting transmission, and sharing costs, shifting electricity from a public utility to a dedicated production material bound to AI data center superparks. This policy orientation based on "self-supplying power" will transform AI data centers from mere power consumers to "power infrastructure investment entities," shifting the demand from just "grid connection capacity" to a full range of power equipment CAPEX, including self-generated power sources, grid connection systems, and on-site distribution systems within the park. This essentially means that the "power-hungry beast" AI will usher in an unprecedented "super bull market" for electric power stocks. During his State of the Union address to Congress, Trump stated that technology companies would be required by the government to construct dedicated power supply systems for their expanding AI infrastructure, rather than drawing additional power from local grids and significantly increasing the load. Chevron Corporation explicitly stated last year that the gas-fired power plant it is building for data centers will initially bypass existing transmission networks to reduce the risk of raising residential electricity prices. Tech companies in the US are now entering the "power grab" era and moving towards the "contract-based self-supplying power" era. The new construction and expansion processes of global AI data centers led by Alphabet Inc. Class C, Microsoft Corporation, and Meta are gaining momentum, highlighting the importance of power supply resources. This is why the investment theme that "the end of AI is electricity" is gaining traction. Furthermore, if the pathway of "self-supplying power" is ultimately institutionalized across the US and other regions like Europe, it will inevitably shift a significant portion of AI capital expenditure to power equipment and grid technology stacks. The global capital markets are pushing the power equipment and grid chain to become the "new mainline" because the AI arms race has expanded the demand from GPU/TPU/AI server computing power to power equipment (gas turbines), transformers/switchgear, transmission and distribution capacity expansion, grid engineering, and dispatching software. From the perspective of power system engineering, "self-supplying power" does not mean operating off-grid but rather commonly involves Behind-the-meter power sources (large gas turbine units, renewables + energy storage, and even nuclear power PPAs) and redundant connections to the public grid. For example, European industrial giant Siemens Energy is clearly benefiting from the demand for gas turbines and power grid equipment driven by AI and has linked its performance and stock price directly to the "AI data center construction wave" driven by North American AI development processes. Similarly, Europe's largest grid operator E.ON is focusing its investments on "preparing for the massive expansion of AI data centers in Europe," and the European utilities sector has become one of the beneficiaries of AI-related market trends.