Training AI Data Center Feed

Meta and Oklo announced an agreement in January 2026 supporting Oklo's plan to develop a 1.2 GW advanced nuclear power campus in Pike County, Ohio. Oklo says Meta's agreement provides a mechanism for power prepayment and funding to advance

Google / AdaniConneX Visakhapatnam AI Hub is a source-backed gigawatt-scale AI infrastructure project in Andhra Pradesh. Google Cloud's April 28, 2026 release says Google broke ground on the AI hub in Visakhapatnam, with the ceremony at Tar

Digital Realty is significantly expanding its UK footprint with a $2 billion "Mega-Campus" project designed to meet the high-density power demands of AI and hyperscale workloads. The campus is projected to reach a total capacity exceeding 3

Vantage Data Centers has significantly expanded its Canadian footprint in early 2026, reaching a projected total capacity of 178 megawatts (MW) across its Quebec portfolio. This expansion is centered in Montreal and Quebec City, with the co

Reliance / Meta Jamnagar AI Data Center is a planned built-to-suit AI-enabled data-center project in Jamnagar, Gujarat. Reliance Industries' June 10, 2026 release says RIL and Meta agreed to develop an AI-enabled data-centre project in Jamn

Yotta D2 / NVIDIA Blackwell AI Supercluster is a planned high-density AI deployment at Yotta's 60 MW D2 data center in the Greater Noida hyperscale campus. Yotta's February 18, 2026 release says the supercluster will deploy 20,736 liquid-co

Mistral's current infrastructure story is Mistral Compute and the Eclairion/Fluidstack Bruyeres-le-Chatel cluster path, not a verified dedicated OVHcloud Paris cluster. Mistral's own Compute page describes a frontier AI cloud with GB200 pro

Stock Farm Road says Fir Hills signed a memorandum of understanding with the Governor of Jeollanam-do Province to develop a proposed AI data center in South Korea. The company describes a project scaling to 3 GW of capacity, with initial si

Vantage Data Centers announced Frontier in August 2025 as a more than $25 billion mega-scale AI campus in Shackelford County, Texas. Vantage says the campus will provide 1.4 GW of GPU compute capacity, sit on 1,200 acres, include 10 data ce

Meta's New Albany, Ohio data-center campus has been active in the community since 2017. Meta's official New Albany page supports more than $1.5 billion of data-center investment in Ohio, more than 300 operational jobs, and renewable-energy

Aurora is an exascale supercomputer at Argonne National Laboratory's Argonne Leadership Computing Facility. ALCF describes Aurora as an Intel and Hewlett Packard Enterprise system launched in January 2025 for large-scale simulation, artific

Meta's Richland Parish Data Center is the company's largest data center project to date in northeast Louisiana. Meta's own Richland Parish page describes a 4 million-square-foot site that will deliver over 2 GW of compute capacity to train

Microsoft Fairwater is the company's Mount Pleasant, Wisconsin AI datacenter campus. Microsoft describes the first Fairwater facility as the world's most powerful AI datacenter, with hundreds of thousands of NVIDIA GPUs, closed-loop liquid

xAI describes Colossus as its gigafactory of compute, built rapidly and doubled to a 200,000 H100 GPU interconnected cluster, with a roadmap to 1 million GPUs. Mississippi state sources announced in January 2026 that xAI is locating the MAC

El Capitan is an exascale supercomputer at Lawrence Livermore National Laboratory for the National Nuclear Security Administration mission. LLNL describes the system as deployed in 2024, and NNSA announced official verification on November

Meta's Eagle Mountain, Utah data center campus is a large-scale facility under development in Utah Valley, approximately 35 miles south of Salt Lake City. The site was selected for its access to renewable energy, land availability, and Utah

Tesla's Dojo supercomputer at the Gigafactory New York (formerly SolarCity factory) in Buffalo is a purpose-built AI training system designed specifically for training Tesla's Autopilot and Full Self-Driving (FSD) neural networks on video d

DeepSeek's AI training cluster in Hangzhou, operated by parent company High-Flyer Quant (幻方科技), gained global attention in early 2025 when its DeepSeek-R1 and V3 models matched GPT-4-class performance at a fraction of the training cost — re

NVIDIA Eos is an AI supercomputer built from 576 NVIDIA DGX H100 systems with 4,608 H100 GPUs. NVIDIA describes the system as delivering 18.4 exaflops of FP8 AI performance for large language models, recommender systems, and other AI worklo

Scaleway Nabu, also described as Nabuchodonosor, is a GPU supercomputer/cluster record located at DC5/PAR2, not a separate 40 MW Vitry-sur-Seine AI data-center facility. Scaleway's October 5, 2023 release says Nabu is an NVIDIA DGX SuperPOD

Frontier is an exascale supercomputer at Oak Ridge National Laboratory's Oak Ridge Leadership Computing Facility. OLCF describes Frontier as delivered in 2021, entering early operations in 2022, and built as an HPE Cray EX system with AMD E

SenseTime's next-generation Artificial Intelligence Data Center in the Lingang New Area of Shanghai commenced operation on January 24, 2022. SenseTime says the AIDC has 3.74 exaFLOPS of designed computing capacity and brings the company's t

Waymo, Alphabet's autonomous vehicle subsidiary, operates large-scale AI training infrastructure for simulation-based learning — training its driving models on billions of miles of simulated driving data generated via its proprietary Waymo

DC5 is Scaleway / OpCore's PAR2 data center in Saint-Ouen-l'Aumone, northwest of Paris. Scaleway describes DC5/PAR2 as powered by renewable wind and hydro energy with GO certification, free and adiabatic cooling, and a PUE of 1.16. Scaleway

Meta's Fort Worth, Texas data center campus is one of the company's primary AI training sites in the United States. The campus, which has expanded multiple times since opening in 2017, hosts dense GPU clusters used to train and fine-tune Me

Meta's Altoona, Iowa data center campus is one of the company's largest and most important AI training facilities in the United States. Located just east of Des Moines in Polk County, the campus has grown substantially since its opening in

Meta's Luleå data center campus in northern Sweden is one of the company's largest and most celebrated facilities globally. Opened in 2013, it was the first Facebook/Meta data center built outside the United States and has grown into a mult

Naver's Gak data center in Chuncheon, Gangwon Province, is South Korea's first and largest hyperscale data center operated by a domestic internet company. Opened in 2013, Gak serves as the primary training and inference site for HyperCLOVA

Microsoft's West Des Moines, Iowa datacenter cluster is a long-running Azure infrastructure footprint that hosted major AI training work for OpenAI. Microsoft says it opened its first West Des Moines datacenter in 2012, added two more campu

ByteDance is developing a massive 500 MW data center campus in Kulai, Johor, to support its global AI research and TikTok workloads outside China. This project, part of a multi-billion dollar regional investment, is specifically engineered

The YTL-NVIDIA AI Cloud in Johor is a landmark 500 MW hyperscale project designed to anchor Malaysia’s position as a global AI hub. Powered by NVIDIA GB200 Grace Blackwell GPUs, the facility is one of the world’s most advanced AI training c

The N01 Campus in Kristiansand, operated by Bulk Data Centers (Bulk Infrastructure Group), is one of Europe’s most strategically positioned hyperscale AI campuses — a purpose-built facility on the southwestern Norwegian coast that combines

EcoDataCenter 2, located in Borlänge, Sweden, is a flagship AI expansion project built on the site of a former paper mill. With an initial capacity of 250 MW and scalability up to 600 MW, it is one of the largest AI-ready campuses in the No

"Stargate Norway" is a $1 billion "AI Gigafactory" located in Kvandal, near Narvik. A strategic collaboration between OpenAI, Nscale, and Aker, the facility is designed to host 100,000 NVIDIA GPUs by late 2026. The site leverages Narvik's A

SINES 4.0 SIN02 is a massive 180 MW expansion of the Start Campus hyperscale site in Sines, Portugal. Evolving into an "AI Gigafactory," the facility is designed to host over 12,000 next-generation NVIDIA GPUs for large-scale AI model train

Magnora ASA, in partnership with Northern Europe Energy Group (NEEG), is developing a 120 MW AI-ready data center in Hämeenlinna, Finland. Located in the MORE Industrial Park, the 150,000-square-meter site is strategically positioned betwee

Microsoft's Markham Megaproject, located on Langstaff Road, is a central pillar of the company's multi-billion dollar infrastructure expansion in Canada. With a planned capacity of 112 MW, the facility is specifically engineered to host hig

Ethiopia's reported National AI Factory is a planned sovereign-AI infrastructure project centered on Addis Ababa and framed as part of the country's broader digital-industrial modernization agenda. Public reporting describes the initiative

Completed in January 2026, Naver’s "B200 4K Cluster" is the largest AI computing cluster in South Korea at launch, featuring 4,000 NVIDIA B200 Blackwell GPUs in a tightly integrated high-density cluster. Located within Naver’s "GAK Sejong"