In reality, falling unit costs will unleash more usage, driving total spend higher. Hyperscale campuses are no longer “lights-out” warehouses. In reality, ongoing costs (maintenance, upgrades, energy and skilled staffing) add up fast and never truly stop. There’s a persistent belief that data centers are mostly a one-time capital expense, but that’s only half the story. PUE benchmarks, redundancy models and site selection factors (such as water and grid proximity) follow entirely new rules.
All told, hyperscalers are planning to spend nearly $700 billion on data center projects in 2026 alone. But as tech companies lined up to report their capex plans for 2026, the rush of data center spending made the figures a lot more interesting — and a lot bigger. “Capital expenditures” are usually a pretty dry metric, referring to a company’s spending on physical assets. In August, Bloomberg reported that the partners were failing to reach consensus. But there were doubts from the beginning, including from Elon Musk, Altman’s business rival, who claimed the project did not have the available funds. Overseeing it all was Trump, who promised to clear away any regulatory hurdles that might slow down the build.
A single large-scale LLM training run can consume 50 million kilowatt-hours, roughly equivalent to the annual electricity use of 5,000 US homes. Goldman Sachs Research (2024) projected that AI data centers could account for 4.5% of total US electricity consumption by 2030, up from under 2% in 2023. Beyond hyperscalers, xAI built a 100,000-GPU H100 cluster in Memphis, Tennessee, in 2024. Microsoft committed $80 billion in FY2025 capital expenditure, largely for data centers. An AI-optimized facility with liquid cooling and high-density GPU racks costs $20 million per MW or more.
Report Your AI Data Center Issue
On September 10, Oracle revealed a five-year, $300 billion deal for compute power, set to begin in 2027. Below, we’ve laid out everything we know about the biggest AI infrastructure projects, including major spending from Meta, Oracle, Microsoft, Google, and OpenAI. It takes a lot of computing power to run an AI product — and as the tech industry races to tap the power of AI models, there’s a parallel race underway to build the infrastructure that will power them. Move from public browsing into CSV, JSON, API, and analyst-grade report workflows when the dataset becomes operational input.
Data Centers Don’t Just Consume Energy
Colocation providers like Equinix, Digital Realty, and Iron Mountain are expanding existing facilities with AI-ready halls, selling capacity to companies that want physical infrastructure without the https://www.electionsscotland.info/why-not-learn-more-about-12/ capital commitment of building from scratch. Beyond the hyperscalers, three other categories are building AI data center capacity. This article explains what makes an AI data center different from a standard one, what the key components are, which companies are building them, what they cost, how they power the AI tools you use every day, and why the energy grid is struggling to keep up. In November 2025, the North American Electric Reliability Corporation warned that building new data centers could negatively affect the electrical grid and cause power outages during extreme weather. Modern hyperscale data centers can exhibit power densities exceeding 100 times those of conventional office buildings, primarily due to the high concentration of servers and cooling systems required to manage continuous digital workloads.
What the data shows about the global buildout.
For higher power density facilities, electricity costs are a dominant operating expense and account for over 10% of the total cost of ownership (TCO) of a data center. The real estate industry, including asset managers, public companies and private investors, has also invested in AI development. Neoclouds such as CoreWeave have gone into debt to buy computer chips from Nvidia for their data centers, and the chips themselves have been used for loan collateral. Citigroup forecasted that $2.8 trillion would be spent on AI data centers by 2030, while McKinsey and Company estimated that almost $7 trillion would be spent globally by that time. Several companies involved in cryptocurrency mining, such as Bitdeer, CoreWeave, Cipher Mining, TeraWulf, IREN, Core Scientific, and CleanSpark have also been involved with AI data centers.
Data Centers Aren’t The Only Driver Of AI’s Resource Demands
For a deeper look at how hyperscalers structure their global data center networks, see our article on what hyperscalers are and how they operate. And national governments, particularly in the EU, Middle East, and Southeast Asia, are funding sovereign AI compute infrastructure to avoid dependency on US hyperscalers. The infrastructure for inference looks more like a distributed cloud than a single facility.
Colocation
As organizations invest more heavily in AI and communities evaluate proposed data center projects, separating perception from reality is becoming more important. The rapid growth of AI infrastructure has raised important questions about energy use, sustainability, security, cost and community impact. The IEA and Goldman Sachs both project that AI-related compute demand will grow faster than any previous technology cycle.
- American Electric Power (AEP), which serves 11 US states, committed $72 billion in November 2025 to add 28 gigawatts of new generation capacity by 2030 specifically to meet data center demand.
- AI accelerators are AI chips used to speed up ML and deep learning (DL) models, natural language processing and other artificial intelligence operations.
- Each layer is specifically designed for AI workloads and cannot simply be borrowed from a standard data center build.
- Every one of them either operates its own AI data centers or has a dedicated partnership with a hyperscaler for exclusive compute access.
11 facilities, 7.4 GW, 20 operators, 5 operational.How many AI data centers are in Germany? 84 facilities, 43.6 GW, 53 operators, 45 operational.How many AI data centers are in France? Your report has been submitted.
- A new 2,250-acre site in Louisiana, dubbed Hyperion, will cost an estimated $10 billion to build out and provide an estimated 5 gigawatts of compute power.
- Self-reporting is the best way we can get this information out to the public!
- The velocity and high computational needs of AI workloads require vast data storage with high-speed memory.
- Power utility companies make upgrades to their infrastructure to handle demands of new data centers, and the price for these changes typically falls on residential or smaller commercial consumers.
- Cooling systems often require substantial water resources, straining local water supplies and ecosystems.
- By July 2026, local community resistance had blocked the construction of AI data centers worth some $130 billion.
- This setup allows businesses to enjoy the benefits of hyperscale, without the major investment.
- It enables better resource usage and flexibility by allowing users to run multiple applications and operating systems on the same physical hardware.
- United States Secretary of Energy Chris Wright expressed support for un-retiring coal plants to power AI data centers.
- This myth fuels community opposition and delays critical AI infrastructure.
- In 2026, major tech companies were estimated to spend $650 billion on AI data centers.
DRI also stated, in areas where there are clusters of data centers, local consumers may end up paying the extra cost of expanding infrastructure. United States Secretary of Energy Chris https://www.softcourier.com/4529/download-exe-password.html Wright expressed support for un-retiring coal plants to power AI data centers. In 2025, the Mountain Valley Pipeline announced plans to expand its capacity by 25% to meet energy needs for data centers. Power utility companies make upgrades to their infrastructure to handle demands of new data centers, and the price for these changes typically falls on residential or smaller commercial consumers.