Hard-core evidence in the industrial chain fights back against “excessive AI computing power”! New Era Energy (NEE.US) joins hands with asset management giants to build a 100 billion dollar data center park

The Zhitong Finance App learned that New Era Energy (NEE.US), the largest power utility company in the US, is cooperating with North American asset management giant Brookfield Asset Management (BAM.US). The project participants plan to jointly invest more than 100 billion US dollars to transform an already closed Cold War-era large-scale uranium enrichment infrastructure into a large-scale data center park supporting large-scale power plants in the US state of Kentucky. The project plan superimposes the latest data center construction curves of tech giants such as Facebook's parent company Meta and Google's parent company Alphabet, to actively catalyze major fundamentals on the global AI computing power theme, which has recently been hit hard, and to empirically refute the “AI excess computing power theory” using hard core evidence from the industrial chain.

The US Department of Energy announced on Monday that the fully privately funded project will include the construction of 2 gigawatts of natural gas power generation facilities at or near the Paducah project site in western Kentucky, as well as up to 2.6 gigawatts of battery energy storage capacity. As a reference, the installed capacity of 1 gigawatt is roughly equivalent to the overall output power of a traditional very large nuclear power plant, and can supply electricity to about 750,000 households at any point in time.

A joint statement from the two sides shows that US power giant New Era Energy and Brookfield Asset Management are cooperating to develop a Kentucky data center and supporting energy project with a total investment of more than 100 billion US dollars, but this indicates the overall private investment scale of the project, which is not equal to average or complete joint investment between the two companies.

When the large-scale data center project was launched, the Trump administration was determined to deal with the surge in data center side power demand brought about by the data center construction frenzy, and this demand is triggering considerable political opposition. Failure to add significant new power capacity limits will threaten one of Trump's key priorities — defeating China in the artificial intelligence race and exacerbating the continued sharp rise in America's already high electricity prices, which could become a political burden as the November midterm elections approach.

Wall Street financial giant Citigroup's latest research report shows that the arms race in the AI era is shifting from “who has the smartest model” to “who can continuously produce intelligence at the lowest cost and with the highest efficiency under physical constraints”: open weight models such as Kimi K3 are rapidly approaching the frontier of closed source, which means that model capabilities accelerate commercialization, but parameter scale, long context, and multi-step agent reasoning expand simultaneously, causing bottlenecks to shift from simple flops to HBM capacity and bandwidth, high-speed GPU interconnection, cluster scheduling, and power access. Nvidia research also points out that when model size, sequence length, and batch expansion, HBM often becomes the main expansion constraint; IEA predictions show that AI data center electricity consumption is growing significantly faster than overall electricity demand, while the power grid construction cycle is generally longer than the data center deployment cycle.

According to a recent research report by senior analyst Brian Nowak from Wall Street financial giant Morgan Stanley, leading the analysis team, the 2027/2028 capital expenditure forecasts for the five largest hyperscale cloud computing and vendors (Meta, Amazon, Microsoft, Google, and SpaceX) in the global market (Meta, Amazon, Microsoft, Google, SpaceX) have been significantly raised again, reaching approximately $1.2 trillion and $1.4 trillion, respectively. The agency's capital expenditure forecast for major US tech giants in 2026 was drastically raised from 433 billion US dollars a year ago to 805 billion US dollars.

From nuclear fuel sites to data center supercities

US Secretary of Energy Chris Wright said that this huge private investment of up to 100 billion US dollars is one of the key measures taken by the US government to increase power generation, create jobs, and ensure America's victory in the artificial intelligence competition.

US Secretary of Energy Chris Wright said in a statement: “The US government is using its assets — such as our federal land — to increase power generation capacity, create jobs, and ensure America wins the AI race.” The Department of Energy said that this private capital investment of over $100 billion is one of the largest investments in Kentucky's history, and is expected to create 8,000 construction jobs and 600 permanent jobs.

According to information, Brookfield Asset Management will develop and operate this 1.8 gigawatt data center campus. The park will occupy part of the U.S. Department of Energy's massive 3,556-acre site; the site was previously used to produce weapons-grade uranium and nuclear reactor fuel. Production activities using today's obsolete uranium enrichment technology known as the gas diffusion method ceased in 2013, and the site is currently in the clean-up phase. The US Department of Energy said construction of the project is expected to be completed in 2031.

The project also includes collaboration with local electric cooperatives. As demand for artificial intelligence tools soars, a global wave of data center expansions has begun. The construction boom has raised concerns across the US about rising water and electricity costs. The Trump administration has been trying to allay these concerns ahead of the midterm elections. One of the measures is to require technology companies to make commitments and bear the associated costs themselves.

Brookfield can currently be described as being a “full-stack infrastructure owner/organizer in the AI computing power era”. In other words, Brookfield is not targeting single-point AI GPU computing power resource leasing, but rather packaging “AI chip computing power supply, server room, AI data center power chain, and underlying energy assets” into an integrated AI era core infrastructure sale/leasing capability, betting that various government organizations and global technology companies will need to obtain the most core AI computing power infrastructure and power resources, such as sky-level AI chips, to win in the ever-upgrading and increasingly intense artificial intelligence competition.

In a statement, Brookfield predicted that global governments or tech giants will require significant capital investment to develop artificial intelligence technology or update and iterate cutting-edge AI models, and that its artificial intelligence infrastructure fund is actively seeking promises from investors to obtain higher returns than its flagship infrastructure funds. Brookfield estimates that achieving a global AI boom will require a capital investment of 7 trillion US dollars, of which AI computing infrastructure will require at least 3 trillion US dollars.

The SoftBank Group, headed by legendary investor Sun Zhengyi, said earlier this year that it is developing a data center park at another former US Department of Energy uranium enrichment site in Ohio, which could be as large as 500 billion US dollars.

100 billion dollar AI power bastion punctures “AI computing power demand collapse theory”

New Era Energy and Brookfield plan to transform the Cold War-era uranium enrichment site in Kentucky into a 1.8 gigawatt AI data center park with about 2 gigawatts of natural gas power generation capacity and maximum 2.6 gigawatts of battery energy storage. The core meaning is not that an additional computer room has been added, but that the AI infrastructure has been upgraded from “large-scale procurement of AI GPU/TPU” to a heavy asset industrial system covering land, independent power supplies, energy storage systems, transmission and distribution, liquid cooling systems, and data center buildings.

2 gigawatts of power generation and 2.6 gigawatts of energy storage cannot be simply added to 4.6 gigawatts of continuous power supply capacity, because the sustainable duration of energy storage has not been disclosed; but this model of “simultaneous power and computing power planning” is the most replicable engineering path for hyperscale parks after public grid access queuing and power supply reliability became the primary bottlenecks. The US Department of Energy has previously listed Paducah as a key site for using federal land to develop AI data centers and supporting power supplies.

This is not an isolated project, but rather resonates strongly with the tech giants' latest construction curves. As of July 29, before the US stock market, Meta's latest official guidance is still capital expenditure of US$125 billion to US$145 billion in 2026, up from US$115 billion to US$135 billion; in July, the company also launched Canada's first 1 gigawatt AI data center, invested more than 13 billion Canadian dollars, and expanded the Louisiana campus to 5 gigawatts of computing capacity and invested more than 50 billion US dollars. Its power plan includes seven new natural gas power generation facilities, three grid-grade batteries, and nuclear power capacity expansion.

Meanwhile, Google's parent company Alphabet raised its 2026 capital expenditure target to US$195 billion to US$205 billion; its second-quarter Google Cloud business (Google Cloud) revenue increased 82% year over year to US$24.768 billion, which directly indicates that demand for AI infrastructure is still accelerating rather than peaking. New Era Energy and Google have also jointly promoted multiple gigawatt parks in the US. The first three projects are under development, and the two sides have put about 3.5 gigawatts of resources into operation or signed contracts.

The project plan, combined with recent data center plans from tech giants such as Facebook's parent company Meta and Google's parent company Alphabet, forms a major fundamental catalyst for the recent hard-hit global AI computing power theme and uses hard-core evidence from the industrial chain to empirically refute the “AI excess computing power theory,” but it is not a sufficient condition for an immediate V-shaped reversal of all AI stocks.

The core of the current AI computing power-themed stock price collapse is no longer a “data center construction project plan,” but rather when huge capital expenses are converted into sustainable free cash flow, and whether debt, financial leasing, and project financing will erode shareholder returns. Alphabet's capital expenditure in the second quarter reached US$44.924 billion, exceeding US$39.069 billion in operating cash flow, resulting in a single quarter's free cash flow of negative US$5.855 billion; the CDS of AI-related companies also expanded markedly, with Meta about 93 basis points, Nvidia about 78 basis points, and the investment-grade CDS index of about 53 basis points. In other words, physical AI computing power orders are still in the supercycle, yet the capital market has moved from “reward investment” to “judging AI return on investment.”

The Kentucky project is scheduled to be completed by 2031, so it is more like a signal covering industrial orders and electricity demand for the next few years, rather than a short-term catalyst that can immediately realize profits for the entire industry chain — the AI computing power industry cycle has not stopped, but the valuation system is changing from “scarce computing power” to “scarce cash flow and capital efficiency.”