[ { "id": "stellantis-mega-campus", "name": "Stellantis Mega Campus", "operator": "Stellantis / Archer Aerospace", "location": "Kokomo, Indiana, USA", "coordinates": { "lat": 40.4864, "lng": -86.1333 }, "status": "Under Construction", "capacity_mw": 100, "gpus": "TBD", "purpose": "AI training and autonomous vehicle research", "announced": "2025-03-15", "expected_online": "2026-Q3", "power_source": "Grid + solar", "investment_usd": 900000000, "sources": [ "https://www.stellantis.com", "https://www.reuters.com" ], "tags": [ "automotive", "training", "autonomous" ], "last_updated": "2026-06-20" }, { "id": "abuden-saudi-ai", "name": "Project Transcendence", "operator": "Saudi Arabia PIF / Google", "location": "Eastern Province, Saudi Arabia", "coordinates": { "lat": 26.3969, "lng": 49.7886 }, "status": "Announced", "capacity_mw": 1000, "gpus": "TBD", "purpose": "Largest AI datacenter outside US — sovereign AI infrastructure", "announced": "2025-05-14", "expected_online": "2027-Q2", "power_source": "Solar + grid", "investment_usd": 100000000000, "sources": [ "https://www.pif.gov.sa", "https://blog.google" ], "tags": [ "sovereign-ai", "training", "middle-east" ], "last_updated": "2026-06-22" }, { "id": "oracle-openai-abl", "name": "Stargate Abilene Campus", "operator": "Oracle / OpenAI / SoftBank", "location": "Abilene, Texas, USA", "coordinates": { "lat": 32.4487, "lng": -99.7331 }, "status": "Phase 1 Operational", "capacity_mw": 300, "gpus": "~100,000 H100s (Phase 1)", "purpose": "Stargate project — $500B OpenAI/Oracle/SoftBank partnership. 5.5 GW total across FIVE sites announced Sep 2025. Abilene is Phase 1.", "announced": "2024-08-01", "expected_online": "2025-Q4 (Phase 1)", "power_source": "Grid + on-site generation", "investment_usd": 500000000000, "sources": [ "https://www.oracle.com", "https://openai.com", "https://www.texastribune.org" ], "tags": [ "training", "frontier-models", "stargate", "mega-project" ], "last_updated": "2026-06-23" }, { "id": "xai-memphis", "name": "xAI Memphis Supercluster", "operator": "xAI (Elon Musk)", "location": "Memphis, Tennessee, USA", "coordinates": { "lat": 35.1175, "lng": -89.9711 }, "status": "Operational", "capacity_mw": 150, "gpus": "100,000 H100s (expanding to 300,000 with H200s/GB200s)", "purpose": "Grok model training — world's largest AI training cluster", "announced": "2024-06-01", "expected_online": "2024-Q4 (operational)", "power_source": "Grid (TVA) + mobile generators", "investment_usd": 3000000000, "sources": [ "https://x.ai", "https://www.commercialappeal.com" ], "tags": [ "training", "frontier-models", "supercluster" ], "last_updated": "2026-06-21" }, { "id": "amazon-cummings-center", "name": "AWS Cummings Center AI Campus", "operator": "Amazon Web Services", "location": "Cummings Research Park, Huntsville, Alabama, USA", "coordinates": { "lat": 34.7334, "lng": -86.6497 }, "status": "Under Construction", "capacity_mw": 200, "gpus": "TBD", "purpose": "AWS AI/ML training and inference infrastructure", "announced": "2025-12-10", "expected_online": "2027-Q1", "power_source": "Grid + nuclear SMR (planned)", "investment_usd": 11000000000, "sources": [ "https://aws.amazon.com", "https://www.al.com" ], "tags": [ "hyperscaler", "training", "inference", "nuclear" ], "last_updated": "2026-06-19" }, { "id": "meta-hyperion-louisiana", "name": "Meta Hyperion", "operator": "Meta Platforms", "location": "Richland Parish, Louisiana, USA", "coordinates": { "lat": 32.4132, "lng": -91.7804 }, "status": "Under Construction", "capacity_mw": 400, "gpus": "TBD (target: 100,000+ GPUs)", "purpose": "Meta's $200B Hyperion project — largest Meta datacenter and one of the largest AI infrastructure projects in history. AI training for Llama and beyond.", "announced": "2024-12-04", "expected_online": "2026-Q4", "power_source": "Grid + 300MW solar", "investment_usd": 200000000000, "sources": [ "https://engineering.fb.com", "https://www.louisianaeconomicdevelopment.com" ], "tags": [ "hyperscaler", "training", "open-source", "hyperion", "mega-project" ], "last_updated": "2026-06-18" }, { "id": "microsoft-mount-pleasant", "name": "Microsoft Mount Pleasant AI Campus", "operator": "Microsoft / OpenAI", "location": "Mount Pleasant, Wisconsin, USA", "coordinates": { "lat": 42.5886, "lng": -87.9235 }, "status": "Under Construction", "capacity_mw": 250, "gpus": "TBD", "purpose": "Stargate Phase 2 — OpenAI training", "announced": "2025-01-20", "expected_online": "2027-Q1", "power_source": "Grid + nuclear (Three Mile Island restart)", "investment_usd": 7000000000, "power_deal": "Constellation Energy — Three Mile Island Unit 1 (837MW)", "sources": [ "https://blogs.microsoft.com", "https://www.constellationenergy.com" ], "on_page_seo": { "h1": "Microsoft Mount Pleasant AI Data Center — Stargate Phase 2", "meta_description": "Track the Microsoft-OpenAI Stargate Phase 2 datacenter in Mount Pleasant, WI: capacity, power, GPU count, and construction timeline.", "keywords": [ "microsoft ai datacenter", "openai stargate", "mount pleasant wisconsin data center", "three mile island ai", "constellation energy ai" ] }, "tags": [ "hyperscaler", "training", "stargate", "nuclear" ], "last_updated": "2026-06-22" }, { "id": "google-bergrath", "name": "Google Bergheim AI Campus", "operator": "Google Cloud", "location": "Bergheim, North Rhine-Westphalia, Germany", "coordinates": { "lat": 50.9753, "lng": 6.6391 }, "status": "Operational (Phase 1)", "capacity_mw": 150, "gpus": "TPU v5p / TPU Trillium", "purpose": "Google DeepMind training + European AI inference", "announced": "2024-01-10", "expected_online": "2025-Q3 (Phase 1 operational)", "power_source": "Grid (100% renewable PPA)", "investment_usd": 1000000000, "sources": [ "https://blog.google", "https://www.google.com/about/datacenters" ], "tags": [ "hyperscaler", "training", "inference", "europe", "tpu" ], "last_updated": "2026-06-20" }, { "id": "blackstone-quintus-x", "name": "Quintus-X AI Campus", "operator": "Blackstone / CoreWeave", "location": "West Columbia, Texas, USA", "coordinates": { "lat": 29.142, "lng": -95.656 }, "status": "Under Construction", "capacity_mw": 350, "gpus": "NVIDIA GB200 NVL72 clusters", "purpose": "CoreWeave cloud GPU infrastructure — leased to AI labs", "announced": "2025-08-20", "expected_online": "2026-Q3", "power_source": "Grid + on-site gas turbines", "investment_usd": 3500000000, "sources": [ "https://www.coreweave.com", "https://www.blackstone.com" ], "tags": [ "neocloud", "inference", "training", "gb200" ], "last_updated": "2026-06-21" }, { "id": "digital-realty-portland", "name": "Digital Realty Portland AI Hub", "operator": "Digital Realty", "location": "Hillsboro, Oregon, USA", "coordinates": { "lat": 45.5228, "lng": -122.9858 }, "status": "Operational", "capacity_mw": 75, "gpus": "Heterogeneous (customer-supplied)", "purpose": "Colocation for AI tenants — carrier-neutral", "announced": "2023-06-15", "power_source": "Grid (largely hydro from BPA)", "investment_usd": 500000000, "sources": [ "https://www.digitalrealty.com" ], "tags": [ "colocation", "inference", "pacific-northwest" ], "last_updated": "2026-06-15" }, { "id": "meta-indiana-ai", "name": "Meta Indiana AI Data Center", "operator": "Meta Platforms", "location": "Jeffersonville, Indiana, USA", "coordinates": { "lat": 38.2856, "lng": -85.7371 }, "status": "Under Construction", "capacity_mw": 200, "gpus": "TBD", "purpose": "$10B AI data center for Llama model training — one of Meta's largest facilities", "announced": "2026-01-15", "expected_online": "2027-Q3", "power_source": "Grid + solar PPA", "investment_usd": 10000000000, "sources": [ "https://engineering.fb.com", "https://www.newsandtribune.com" ], "tags": [ "hyperscaler", "training", "open-source", "midwest" ], "last_updated": "2026-06-23" }, { "id": "missouri-hyperscale-corridor", "name": "Missouri Hyperscale Corridor", "operator": "Multiple (Google, Meta, LLC operators)", "location": "Kansas City / Independence, Missouri, USA", "coordinates": { "lat": 39.0997, "lng": -94.5786 }, "status": "Announced", "capacity_mw": 500, "gpus": "TBD", "purpose": "Emerging hyperscale cluster — multiple operators attracted by grid capacity and incentives", "announced": "2025-11-01", "expected_online": "2027-Q4", "power_source": "Grid (Evergy) + planned solar", "investment_usd": 5000000000, "sources": [ "https://www.kcur.org", "https://www.missourieconomicdevelopment.com" ], "tags": [ "hyperscaler", "training", "midwest", "emerging-cluster" ], "last_updated": "2026-06-23" }, { "id": "google-the-dalles", "name": "Google The Dalles AI Campus", "operator": "Google Cloud", "location": "The Dalles, Oregon, USA", "coordinates": { "lat": 45.6045, "lng": -121.1858 }, "status": "Operational (expanding)", "capacity_mw": 225, "gpus": "TPU v5p / TPU Trillium / H100 clusters", "purpose": "Google DeepMind training — one of Google's original and largest AI campuses", "announced": "2023-10-01", "expected_online": "Operational (Phase 4 expansion 2026-Q4)", "power_source": "Grid (BPA hydro) + solar PPA", "investment_usd": 600000000, "sources": [ "https://www.google.com/about/datacenters", "https://www.bpa.gov" ], "tags": [ "hyperscaler", "training", "tpu", "pacific-northwest", "operational" ], "last_updated": "2026-06-22" }, { "id": "crusoe-western-texas", "name": "Crusoe West Texas AI Campus", "operator": "Crusoe Energy Systems", "location": "Monahans, Texas, USA", "coordinates": { "lat": 31.5835, "lng": -102.8954 }, "status": "Phase 1 Operational", "capacity_mw": 100, "gpus": "NVIDIA H100s (expanding to GB200s)", "purpose": "Flared-gas-powered AI compute — first commercial stranded-energy AI campus", "announced": "2024-03-20", "expected_online": "2025-Q3 (Phase 1 operational)", "power_source": "Stranded natural gas (on-site generators)", "investment_usd": 800000000, "sources": [ "https://crusoeenergy.com", "https://www.reuters.com" ], "tags": [ "neocloud", "training", "stranded-energy", "texas", "innovative" ], "last_updated": "2026-06-20" }, { "id": "microsoft-arizona-mesa", "name": "Microsoft Mesa AI Campus", "operator": "Microsoft Azure", "location": "Mesa, Arizona, USA", "coordinates": { "lat": 33.4152, "lng": -111.8315 }, "status": "Under Construction", "capacity_mw": 180, "gpus": "TBD", "purpose": "Azure AI training and inference — part of Microsoft's $80B AI infrastructure plan", "announced": "2025-09-05", "expected_online": "2027-Q2", "power_source": "Grid (SRP) + solar PPA", "investment_usd": 2500000000, "sources": [ "https://blogs.microsoft.com", "https://www.srpnet.com" ], "tags": [ "hyperscaler", "training", "inference", "arizona", "solar" ], "last_updated": "2026-06-19" }, { "id": "microsoft-quincy-wa", "name": "Microsoft Quincy Data Center Campus", "operator": "Microsoft Azure", "location": "Quincy, Washington, USA", "coordinates": { "lat": 47.2349, "lng": -119.853 }, "status": "Operational", "capacity_mw": 200, "gpus": "H100 / H200 clusters (Azure ND series)", "purpose": "Azure AI training and inference — one of Microsoft's original hyperscale campuses powered by hydroelectric", "announced": "2007-01-01", "expected_online": "Operational (multiple phases, ongoing expansion)", "power_source": "Grid (Grant County PUD hydro)", "investment_usd": 2000000000, "sources": [ "https://blogs.microsoft.com", "https://www.google.com/about/datacenters" ], "tags": [ "hyperscaler", "training", "inference", "pacific-northwest", "hydro", "operational" ], "last_updated": "2026-06-24" }, { "id": "microsoft-san-antonio-tx", "name": "Microsoft San Antonio Data Center", "operator": "Microsoft Azure", "location": "San Antonio, Texas, USA", "coordinates": { "lat": 29.4241, "lng": -98.4936 }, "status": "Operational", "capacity_mw": 160, "gpus": "H100 / H200 clusters", "purpose": "Azure AI infrastructure — large-scale training and inference for South-Central US region", "announced": "2008-01-01", "expected_online": "Operational (expanding)", "power_source": "Grid (CPS Energy) + wind PPA", "investment_usd": 1500000000, "sources": [ "https://blogs.microsoft.com", "https://www.cpsenergy.com" ], "tags": [ "hyperscaler", "training", "inference", "texas", "wind", "operational" ], "last_updated": "2026-06-24" }, { "id": "microsoft-west-des-moines-ia", "name": "Microsoft West Des Moines AI Campus", "operator": "Microsoft Azure / OpenAI", "location": "West Des Moines, Iowa, USA", "coordinates": { "lat": 41.5775, "lng": -93.7114 }, "status": "Operational (expanding)", "capacity_mw": 250, "gpus": "H100 / H200 / GB200 clusters", "purpose": "OpenAI training infrastructure — one of the largest single-site AI training facilities in the US", "announced": "2009-01-01", "expected_online": "Operational (Phase 5 expansion 2027-Q1)", "power_source": "Grid (MidAmerican Energy) + wind", "investment_usd": 4000000000, "sources": [ "https://blogs.microsoft.com", "https://www.midamericanenergy.com" ], "tags": [ "hyperscaler", "training", "openai", "midwest", "wind", "operational" ], "last_updated": "2026-06-24" }, { "id": "microsoft-boydton-va", "name": "Microsoft Boydton Data Center", "operator": "Microsoft Azure", "location": "Boydton, Virginia, USA", "coordinates": { "lat": 36.6658, "lng": -78.3875 }, "status": "Operational", "capacity_mw": 180, "gpus": "H100 / L40S clusters", "purpose": "Azure AI inference and training — East US region hub in Northern Virginia data center alley", "announced": "2010-01-01", "expected_online": "Operational (expanding)", "power_source": "Grid (Dominion Energy) + solar PPAs", "investment_usd": 1800000000, "sources": [ "https://blogs.microsoft.com", "https://www.dominionenergy.com" ], "tags": [ "hyperscaler", "training", "inference", "virginia", "east-us", "operational" ], "last_updated": "2026-06-24" }, { "id": "microsoft-cheyenne-wy", "name": "Microsoft Cheyenne Data Center", "operator": "Microsoft Azure", "location": "Cheyenne, Wyoming, USA", "coordinates": { "lat": 41.14, "lng": -104.8202 }, "status": "Under Construction", "capacity_mw": 150, "gpus": "TBD (target: H200 / GB200)", "purpose": "Azure AI training — low-latency central US region with abundant wind power", "announced": "2025-06-15", "expected_online": "2027-Q1", "power_source": "Grid (PacifiCorp) + wind PPA", "investment_usd": 1200000000, "sources": [ "https://blogs.microsoft.com", "https://www.wyomingbusiness.org" ], "tags": [ "hyperscaler", "training", "wyoming", "wind" ], "last_updated": "2026-06-24" }, { "id": "google-council-bluffs-ia", "name": "Google Council Bluffs Data Center", "operator": "Google Cloud", "location": "Council Bluffs, Iowa, USA", "coordinates": { "lat": 41.2619, "lng": -95.8608 }, "status": "Operational (expanding)", "capacity_mw": 200, "gpus": "TPU v5p / TPU Trillium / H100 clusters", "purpose": "Google Cloud AI training and inference — one of Google's largest US campuses", "announced": "2007-01-01", "expected_online": "Operational (ongoing expansion)", "power_source": "Grid (MidAmerican Energy) + wind", "investment_usd": 5000000000, "sources": [ "https://www.google.com/about/datacenters" ], "tags": [ "hyperscaler", "training", "inference", "tpu", "midwest", "wind", "operational" ], "last_updated": "2026-06-24" }, { "id": "google-douglas-county-ga", "name": "Google Douglas County Data Center", "operator": "Google Cloud", "location": "Douglas County, Georgia, USA", "coordinates": { "lat": 33.7549, "lng": -84.7607 }, "status": "Operational", "capacity_mw": 175, "gpus": "TPU v5p / TPU Trillium / H100 clusters", "purpose": "Google Cloud AI infrastructure — Southeast US region hub", "announced": "2007-01-01", "expected_online": "Operational (expanding)", "power_source": "Grid (Georgia Power) + solar PPAs", "investment_usd": 2500000000, "sources": [ "https://www.google.com/about/datacenters" ], "tags": [ "hyperscaler", "training", "inference", "tpu", "southeast-us", "solar", "operational" ], "last_updated": "2026-06-24" }, { "id": "google-mayes-county-ok", "name": "Google Mayes County Data Center", "operator": "Google Cloud", "location": "Pryor Creek, Mayes County, Oklahoma, USA", "coordinates": { "lat": 36.302, "lng": -95.242 }, "status": "Operational", "capacity_mw": 150, "gpus": "TPU v5e / TPU Trillium / H100 clusters", "purpose": "Google Cloud AI training and inference — central US region", "announced": "2010-01-01", "expected_online": "Operational (expanding)", "power_source": "Grid (GRDA) + wind", "investment_usd": 3000000000, "sources": [ "https://www.google.com/about/datacenters" ], "tags": [ "hyperscaler", "training", "inference", "tpu", "oklahoma", "wind", "operational" ], "last_updated": "2026-06-24" }, { "id": "google-new-albany-oh", "name": "Google New Albany Data Center", "operator": "Google Cloud", "location": "New Albany, Ohio, USA", "coordinates": { "lat": 40.0812, "lng": -82.8185 }, "status": "Operational (expanding)", "capacity_mw": 160, "gpus": "TPU v5p / TPU Trillium / H100 clusters", "purpose": "Google Cloud AI infrastructure — Midwest US region hub", "announced": "2019-01-01", "expected_online": "Operational (Phase 3 expansion 2027-Q2)", "power_source": "Grid (AEP Ohio) + solar PPAs", "investment_usd": 2000000000, "sources": [ "https://www.google.com/about/datacenters" ], "tags": [ "hyperscaler", "training", "inference", "tpu", "midwest", "solar", "operational" ], "last_updated": "2026-06-24" }, { "id": "google-storey-county-nv", "name": "Google Storey County Data Center", "operator": "Google Cloud", "location": "Storey County, Nevada, USA", "coordinates": { "lat": 39.52, "lng": -119.75 }, "status": "Under Construction", "capacity_mw": 180, "gpus": "TPU Trillium / GB200 clusters", "purpose": "Google Cloud AI training — Tahoe-Reno Industrial Center campus", "announced": "2025-03-10", "expected_online": "2027-Q1", "power_source": "Grid (NV Energy) + geothermal PPA (Fervo Energy)", "investment_usd": 2200000000, "sources": [ "https://www.google.com/about/datacenters", "https://fervoenergy.com" ], "tags": [ "hyperscaler", "training", "tpu", "nevada", "geothermal" ], "last_updated": "2026-06-24" }, { "id": "amazon-ashburn-va", "name": "AWS Ashburn Data Center Complex", "operator": "Amazon Web Services", "location": "Ashburn, Virginia, USA", "coordinates": { "lat": 39.0438, "lng": -77.4875 }, "status": "Operational", "capacity_mw": 300, "gpus": "H100 / H200 / L40S / Trainium clusters", "purpose": "AWS US-East-1 region — largest AWS data center cluster, primary AI training and inference hub", "announced": "2006-01-01", "expected_online": "Operational (continuous expansion)", "power_source": "Grid (Dominion Energy) + solar PPAs", "investment_usd": 5000000000, "sources": [ "https://aws.amazon.com", "https://www.dominionenergy.com" ], "tags": [ "hyperscaler", "training", "inference", "virginia", "east-us", "operational" ], "last_updated": "2026-06-24" }, { "id": "amazon-boardman-or", "name": "AWS Boardman Data Center", "operator": "Amazon Web Services", "location": "Boardman, Oregon, USA", "coordinates": { "lat": 45.839, "lng": -119.726 }, "status": "Operational", "capacity_mw": 120, "gpus": "H100 / Trainium clusters", "purpose": "AWS US-West-2 region — Pacific Northwest AI training hub powered by hydro", "announced": "2011-01-01", "expected_online": "Operational (expanding)", "power_source": "Grid (BPA hydro) + solar", "investment_usd": 1500000000, "sources": [ "https://aws.amazon.com", "https://www.bpa.gov" ], "tags": [ "hyperscaler", "training", "inference", "pacific-northwest", "hydro", "operational" ], "last_updated": "2026-06-24" }, { "id": "amazon-umatilla-or", "name": "AWS Umatilla Data Center", "operator": "Amazon Web Services", "location": "Umatilla, Oregon, USA", "coordinates": { "lat": 45.808, "lng": -119.336 }, "status": "Under Construction", "capacity_mw": 200, "gpus": "H200 / Trainium2 / GB200 clusters", "purpose": "AWS US-West-2 expansion — large-scale AI training campus", "announced": "2025-04-20", "expected_online": "2027-Q2", "power_source": "Grid (BPA hydro) + wind", "investment_usd": 2500000000, "sources": [ "https://aws.amazon.com", "https://www.bpa.gov" ], "tags": [ "hyperscaler", "training", "inference", "pacific-northwest", "hydro", "wind" ], "last_updated": "2026-06-24" }, { "id": "meta-eagle-mountain-ut", "name": "Meta Eagle Mountain Data Center", "operator": "Meta Platforms", "location": "Eagle Mountain, Utah, USA", "coordinates": { "lat": 40.311, "lng": -111.783 }, "status": "Operational (expanding)", "capacity_mw": 150, "gpus": "H100 / MI300X / MTIA clusters", "purpose": "Meta AI training for Llama models — one of Meta's largest campuses", "announced": "2018-01-01", "expected_online": "Operational (Phase 3 expansion 2026-Q4)", "power_source": "Grid (Rocky Mountain Power) + solar", "investment_usd": 1800000000, "sources": [ "https://engineering.fb.com" ], "tags": [ "hyperscaler", "training", "open-source", "utah", "solar", "operational" ], "last_updated": "2026-06-24" }, { "id": "meta-dekalb-il", "name": "Meta DeKalb Data Center", "operator": "Meta Platforms", "location": "DeKalb, Illinois, USA", "coordinates": { "lat": 41.929, "lng": -88.749 }, "status": "Operational", "capacity_mw": 120, "gpus": "H100 / MI300X clusters", "purpose": "Meta AI training — Midwest region campus for Llama and recommendation models", "announced": "2020-01-01", "expected_online": "Operational (expanding)", "power_source": "Grid (ComEd) + 100% renewable match", "investment_usd": 1200000000, "sources": [ "https://engineering.fb.com" ], "tags": [ "hyperscaler", "training", "open-source", "midwest", "operational" ], "last_updated": "2026-06-24" }, { "id": "meta-sarpy-county-ne", "name": "Meta Sarpy County Data Center", "operator": "Meta Platforms", "location": "Papillion, Sarpy County, Nebraska, USA", "coordinates": { "lat": 41.154, "lng": -96.048 }, "status": "Operational", "capacity_mw": 130, "gpus": "H100 / MI300X clusters", "purpose": "Meta AI infrastructure — central US region campus", "announced": "2020-01-01", "expected_online": "Operational (expanding)", "power_source": "Grid (OPPD) + wind", "investment_usd": 1500000000, "sources": [ "https://engineering.fb.com" ], "tags": [ "hyperscaler", "training", "open-source", "midwest", "wind", "operational" ], "last_updated": "2026-06-24" }, { "id": "meta-temple-tx", "name": "Meta Temple Data Center", "operator": "Meta Platforms", "location": "Temple, Texas, USA", "coordinates": { "lat": 31.0982, "lng": -97.3428 }, "status": "Under Construction", "capacity_mw": 180, "gpus": "H200 / MI300X / GB200 clusters (planned)", "purpose": "Meta AI training — $800M campus expansion for Llama and Reality Labs workloads", "announced": "2024-10-01", "expected_online": "2027-Q1", "power_source": "Grid (Oncor) + solar PPA", "investment_usd": 800000000, "sources": [ "https://engineering.fb.com", "https://www.templetx.gov" ], "tags": [ "hyperscaler", "training", "open-source", "texas", "solar" ], "last_updated": "2026-06-24" }, { "id": "meta-huntsville-al", "name": "Meta Huntsville Data Center", "operator": "Meta Platforms", "location": "Huntsville, Alabama, USA", "coordinates": { "lat": 34.68, "lng": -86.74 }, "status": "Operational", "capacity_mw": 140, "gpus": "H100 / MI300X clusters", "purpose": "Meta AI infrastructure — Southeast US campus for training and inference", "announced": "2018-01-01", "expected_online": "Operational (expanding)", "power_source": "Grid (TVA) + solar", "investment_usd": 1500000000, "sources": [ "https://engineering.fb.com" ], "tags": [ "hyperscaler", "training", "open-source", "southeast-us", "solar", "operational" ], "last_updated": "2026-06-24" }, { "id": "oracle-mesa-az", "name": "Oracle Mesa Data Center Campus", "operator": "Oracle Cloud", "location": "Mesa, Arizona, USA", "coordinates": { "lat": 33.37, "lng": -111.72 }, "status": "Under Construction", "capacity_mw": 200, "gpus": "H200 / B200 / GB200 clusters", "purpose": "Oracle Cloud AI infrastructure — OCI supercluster training for OpenAI and other AI labs", "announced": "2025-07-15", "expected_online": "2027-Q1", "power_source": "Grid (SRP) + solar PPA", "investment_usd": 3000000000, "sources": [ "https://www.oracle.com", "https://www.srpnet.com" ], "tags": [ "hyperscaler", "training", "inference", "arizona", "solar", "stargate" ], "last_updated": "2026-06-24" }, { "id": "oracle-west-jordan-ut", "name": "Oracle West Jordan Data Center", "operator": "Oracle Cloud", "location": "West Jordan, Utah, USA", "coordinates": { "lat": 40.61, "lng": -111.94 }, "status": "Under Construction", "capacity_mw": 150, "gpus": "B200 / GB200 clusters (planned)", "purpose": "Oracle Cloud Infrastructure — AI supercluster for OCI compute services", "announced": "2025-10-01", "expected_online": "2027-Q2", "power_source": "Grid (Rocky Mountain Power) + solar", "investment_usd": 1800000000, "sources": [ "https://www.oracle.com" ], "tags": [ "hyperscaler", "training", "inference", "utah", "solar" ], "last_updated": "2026-06-24" }, { "id": "coreweave-nj", "name": "CoreWeave New Jersey AI Campus", "operator": "CoreWeave", "location": "Edison, New Jersey, USA", "coordinates": { "lat": 40.5187, "lng": -74.4121 }, "status": "Operational", "capacity_mw": 80, "gpus": "H100 / H200 / L40S clusters", "purpose": "CoreWeave GPU cloud — primary East Coast neocloud facility", "announced": "2022-01-01", "expected_online": "Operational (expanding)", "power_source": "Grid (PSE&G)", "investment_usd": 500000000, "sources": [ "https://www.coreweave.com" ], "tags": [ "neocloud", "training", "inference", "east-us", "operational" ], "last_updated": "2026-06-24" }, { "id": "coreweave-illinois", "name": "CoreWeave Illinois AI Campus", "operator": "CoreWeave", "location": "Aurora, Illinois, USA", "coordinates": { "lat": 41.7606, "lng": -88.1538 }, "status": "Under Construction", "capacity_mw": 120, "gpus": "GB200 NVL72 / H200 clusters", "purpose": "CoreWeave GPU cloud — Midwest hub for AI lab leasing", "announced": "2025-05-01", "expected_online": "2026-Q4", "power_source": "Grid (ComEd) + on-site generation", "investment_usd": 800000000, "sources": [ "https://www.coreweave.com" ], "tags": [ "neocloud", "training", "inference", "midwest", "gb200" ], "last_updated": "2026-06-24" }, { "id": "lambda-labs-cloud", "name": "Lambda Cloud GPU Infrastructure", "operator": "Lambda Labs", "location": "San Jose, California, USA", "coordinates": { "lat": 37.3382, "lng": -121.8947 }, "status": "Operational", "capacity_mw": 50, "gpus": "H100 / H200 / GB200 clusters", "purpose": "Lambda on-demand GPU cloud — AI training and inference for startups and researchers", "announced": "2023-01-01", "expected_online": "Operational (expanding)", "power_source": "Grid (PG&E)", "investment_usd": 300000000, "sources": [ "https://lambdalabs.com" ], "tags": [ "neocloud", "training", "inference", "california", "operational" ], "last_updated": "2026-06-24" }, { "id": "together-ai-infra", "name": "Together AI Training Cluster", "operator": "Together AI", "location": "San Francisco, California, USA", "coordinates": { "lat": 37.7749, "lng": -122.4194 }, "status": "Operational", "capacity_mw": 30, "gpus": "H100 / H200 clusters (partner-hosted)", "purpose": "Together AI inference and fine-tuning platform — GPU infrastructure for open-source model hosting", "announced": "2023-06-01", "expected_online": "Operational", "power_source": "Grid (partner data centers)", "investment_usd": 200000000, "sources": [ "https://www.together.ai" ], "tags": [ "neocloud", "inference", "training", "california", "open-source" ], "last_updated": "2026-06-24" }, { "id": "crusoe-north-dakota", "name": "Crusoe North Dakota Digital Flaring Campus", "operator": "Crusoe Energy Systems", "location": "Bakken Formation, North Dakota, USA", "coordinates": { "lat": 46.8083, "lng": -100.7837 }, "status": "Phase 1 Operational", "capacity_mw": 50, "gpus": "H100 / L40S clusters", "purpose": "Stranded-gas-powered AI compute — reducing flaring emissions while powering GPU clusters", "announced": "2023-01-01", "expected_online": "Operational (Phase 2 expanding)", "power_source": "Stranded natural gas (on-site generators)", "investment_usd": 300000000, "sources": [ "https://crusoeenergy.com" ], "tags": [ "neocloud", "training", "stranded-energy", "north-dakota", "innovative", "operational" ], "last_updated": "2026-06-24" }, { "id": "uae-g42-stargate", "name": "UAE G42 Stargate Abu Dhabi", "operator": "G42 / Microsoft / OpenAI", "location": "Abu Dhabi, UAE", "coordinates": { "lat": 24.4539, "lng": 54.3773 }, "status": "Announced", "capacity_mw": 500, "gpus": "TBD (target: GB200 / Rubin)", "purpose": "Sovereign AI infrastructure — UAE's flagship AI campus for the Middle East and Africa region", "announced": "2025-04-15", "expected_online": "2027-Q4", "power_source": "Grid + nuclear (Barakah) + solar", "investment_usd": 20000000000, "sources": [ "https://www.g42.ai", "https://blogs.microsoft.com" ], "tags": [ "sovereign-ai", "training", "middle-east", "nuclear", "mega-project" ], "last_updated": "2026-06-24" }, { "id": "saudi-neom-ai", "name": "NEOM AI Campus", "operator": "NEOM / Saudi PIF", "location": "NEOM, Tabuk Region, Saudi Arabia", "coordinates": { "lat": 28.0333, "lng": 35.0333 }, "status": "Announced", "capacity_mw": 300, "gpus": "TBD", "purpose": "Futuristic AI campus within NEOM megaproject — sovereign AI and smart city compute", "announced": "2025-09-01", "expected_online": "2028-Q2", "power_source": "Solar + wind (NEOM renewable grid)", "investment_usd": 5000000000, "sources": [ "https://www.neom.com", "https://www.pif.gov.sa" ], "tags": [ "sovereign-ai", "training", "middle-east", "solar", "wind", "mega-project" ], "last_updated": "2026-06-24" }, { "id": "south-korea-ai", "name": "South Korea AI Infrastructure Cluster", "operator": "Samsung / Naver Cloud", "location": "Gyeonggi Province, South Korea", "coordinates": { "lat": 37.2636, "lng": 127.0286 }, "status": "Under Construction", "capacity_mw": 100, "gpus": "Samsung SAIDIA / H100 / custom accelerators", "purpose": "South Korea's national AI infrastructure — sovereign AI training for HyperCLOVA X and government", "announced": "2024-09-01", "expected_online": "2027-Q1", "power_source": "Grid (KEPCO) + nuclear", "investment_usd": 3000000000, "sources": [ "https://www.samsung.com", "https://www.naver.com" ], "tags": [ "sovereign-ai", "training", "asia", "nuclear" ], "last_updated": "2026-06-24" }, { "id": "japan-sakura-internet", "name": "Sakura Internet Ishikari AI Campus", "operator": "Sakura Internet", "location": "Ishikari, Hokkaido, Japan", "coordinates": { "lat": 43.227, "lng": 141.322 }, "status": "Phase 1 Operational", "capacity_mw": 50, "gpus": "H100 / H200 / GB200 clusters", "purpose": "Japan's leading AI cloud — powered by Hokkaido's cool climate and renewable energy", "announced": "2023-12-01", "expected_online": "2026-Q4 (Phase 2)", "power_source": "Grid (HEPCO) + geothermal + wind", "investment_usd": 800000000, "sources": [ "https://www.sakura.ad.jp", "https://www.meti.go.jp" ], "tags": [ "neocloud", "training", "inference", "asia", "geothermal", "innovative" ], "last_updated": "2026-06-24" }, { "id": "india-yotta-data", "name": "Yotta D1 AI Data Center", "operator": "Yotta Infrastructure / Hiranandani Group", "location": "Navi Mumbai, Maharashtra, India", "coordinates": { "lat": 19.085, "lng": 73.02 }, "status": "Operational", "capacity_mw": 50, "gpus": "H100 / H200 clusters (scaling to 25,000+ GPUs)", "purpose": "India's largest AI data center — sovereign AI infrastructure for IndiaAI mission", "announced": "2024-01-01", "expected_online": "Operational (expanding to 500MW campus)", "power_source": "Grid (MSEDCL) + solar", "investment_usd": 500000000, "sources": [ "https://www.yottadatacenter.com", "https://www.meity.gov.in" ], "tags": [ "sovereign-ai", "training", "inference", "asia", "solar", "operational" ], "last_updated": "2026-06-24" }, { "id": "france-jean-zay", "name": "Jean Zay Supercomputer Expansion", "operator": "GENCI / IDRIS (French Government)", "location": "Saclay, Île-de-France, France", "coordinates": { "lat": 48.7, "lng": 2.2 }, "status": "Operational (expanding)", "capacity_mw": 15, "gpus": "H100 / H200 clusters (2,048+ GPUs)", "purpose": "France's national AI supercomputer — EuroHPC joint undertaking, training for French and European AI research", "announced": "2023-06-01", "expected_online": "Operational (expansion 2026-Q4)", "power_source": "Grid (EDF, 90% nuclear)", "investment_usd": 250000000, "sources": [ "https://www.genci.fr", "https://www.idris.fr" ], "tags": [ "sovereign-ai", "training", "europe", "nuclear", "supercomputer", "operational" ], "last_updated": "2026-06-24" }, { "id": "uk-ai-research-resource", "name": "UK AI Research Resource (AIRR)", "operator": "UKRI / University of Bristol", "location": "Bristol, United Kingdom", "coordinates": { "lat": 51.4545, "lng": -2.5879 }, "status": "Operational", "capacity_mw": 20, "gpus": "H100 clusters (Isambard-AI, 5,448 GPUs)", "purpose": "UK's national AI supercomputer — Isambard-AI at Bristol, for academic and industry AI research", "announced": "2023-10-01", "expected_online": "Operational", "power_source": "Grid + on-site solar", "investment_usd": 300000000, "sources": [ "https://www.ukri.org", "https://www.bristol.ac.uk" ], "tags": [ "sovereign-ai", "training", "europe", "supercomputer", "operational" ], "last_updated": "2026-06-24" }, { "id": "canada-toronto-ai", "name": "Canada AI Compute Infrastructure (Toronto)", "operator": "Compute Canada / Vector Institute", "location": "Toronto, Ontario, Canada", "coordinates": { "lat": 43.6532, "lng": -79.3832 }, "status": "Operational", "capacity_mw": 15, "gpus": "H100 / H200 clusters", "purpose": "Canada's national AI compute — supporting Vector Institute and academic AI research", "announced": "2023-09-01", "expected_online": "Operational (expanding)", "power_source": "Grid (hydro + nuclear, 90% clean)", "investment_usd": 200000000, "sources": [ "https://www.computecanada.ca", "https://vectorinstitute.ai" ], "tags": [ "sovereign-ai", "training", "north-america", "hydro", "nuclear", "operational" ], "last_updated": "2026-06-24" }, { "id": "singapore-ai", "name": "Singapore AI Compute Cluster", "operator": "Singtel / Digital Realty", "location": "Singapore", "coordinates": { "lat": 1.3521, "lng": 103.8198 }, "status": "Under Construction", "capacity_mw": 60, "gpus": "H100 / H200 / GB200 clusters", "purpose": "Singapore's national AI infrastructure — Southeast Asia AI hub for sovereign and regional AI compute", "announced": "2025-02-01", "expected_online": "2027-Q1", "power_source": "Grid (SP Group) + solar (rooftop + import)", "investment_usd": 800000000, "sources": [ "https://www.singtel.com", "https://www.digitalrealty.com" ], "tags": [ "sovereign-ai", "training", "inference", "asia", "solar" ], "last_updated": "2026-06-24" }, { "id": "israel-nvidia-tel-aviv", "name": "NVIDIA Israel-1 Supercomputer", "operator": "NVIDIA", "location": "Tel Aviv, Israel", "coordinates": { "lat": 32.0853, "lng": 34.7818 }, "status": "Operational", "capacity_mw": 10, "gpus": "H100 / H200 / GH200 clusters", "purpose": "NVIDIA's Israel-1 supercomputer — AI research and development hub, one of the fastest in the region", "announced": "2023-05-01", "expected_online": "Operational", "power_source": "Grid (IEC) + solar", "investment_usd": 150000000, "sources": [ "https://blogs.nvidia.com" ], "tags": [ "nvidia", "training", "inference", "middle-east", "supercomputer", "operational" ], "last_updated": "2026-06-24" }, { "id": "tmi-nuclear-powered", "name": "Three Mile Island Nuclear AI Campus", "operator": "Microsoft / Constellation Energy", "location": "Middletown, Pennsylvania, USA", "coordinates": { "lat": 40.154, "lng": -76.724 }, "status": "Announced", "capacity_mw": 837, "gpus": "TBD (dedicated to Microsoft / OpenAI training)", "purpose": "Nuclear-powered AI datacenter — 20-year PPA to restart TMI Unit 1 for Microsoft AI infrastructure", "announced": "2024-09-20", "expected_online": "2028-Q1 (pending NRC approval)", "power_source": "Nuclear (Three Mile Island Unit 1 restart, 837MW)", "investment_usd": 1600000000, "power_deal": "Constellation Energy — 20-year PPA for 837MW", "sources": [ "https://www.constellationenergy.com", "https://blogs.microsoft.com" ], "tags": [ "nuclear", "training", "innovative", "pennsylvania", "mega-project" ], "last_updated": "2026-06-24" }, { "id": "quad-cities-nuclear", "name": "Quad Cities Nuclear AI Power Deal", "operator": "Microsoft / Constellation Energy", "location": "Cordova, Illinois, USA", "coordinates": { "lat": 41.67, "lng": -90.38 }, "status": "Announced", "capacity_mw": 828, "gpus": "TBD", "purpose": "Nuclear power restart for AI datacenter supply — Quad Cities nuclear station PPA (Units 1 & 2, ~1870MW combined)", "announced": "2025-09-01", "expected_online": "2028-Q2 (pending regulatory approval)", "power_source": "Nuclear (Quad Cities Units 1 & 2 restart, ~1870MW combined)", "investment_usd": 1000000000, "power_deal": "Constellation Energy — Quad Cities nuclear PPA", "sources": [ "https://www.constellationenergy.com", "https://blogs.microsoft.com" ], "tags": [ "nuclear", "training", "innovative", "illinois", "mega-project" ], "last_updated": "2026-06-24" }, { "id": "fervo-geothermal", "name": "Fervo Energy Geothermal AI Campus", "operator": "Fervo Energy / Google", "location": "northern Nevada, USA", "coordinates": { "lat": 39.85, "lng": -117.55 }, "status": "Phase 1 Operational", "capacity_mw": 50, "gpus": "TPU v5p / TPU Trillium clusters", "purpose": "Next-gen geothermal-powered AI compute — enhanced geothermal systems (EGS) powering Google datacenters", "announced": "2023-08-01", "expected_online": "Operational (scaling to 350MW by 2027)", "power_source": "Enhanced geothermal (Fervo Energy EGS, 24/7 carbon-free)", "investment_usd": 400000000, "sources": [ "https://fervoenergy.com", "https://blog.google" ], "tags": [ "geothermal", "training", "innovative", "nevada", "renewable", "operational" ], "last_updated": "2026-06-24" }, { "id": "subsea-datacenter", "name": "Project Natick Subsea Datacenter (Concept)", "operator": "Microsoft Research", "location": "Orkney, Scotland, UK", "coordinates": { "lat": 58.7, "lng": -3.3 }, "status": "Announced", "capacity_mw": 5, "gpus": "TBD (experimental)", "purpose": "Experimental underwater datacenter concept — natural cooling and reduced latency for coastal populations", "announced": "2018-06-01", "expected_online": "TBD (research phase, potential commercial deployment 2028+)", "power_source": "Grid (subsea cable) + tidal (concept)", "investment_usd": 50000000, "sources": [ "https://docs.microsoft.com/en-us/sustainability", "https://blogs.microsoft.com" ], "tags": [ "innovative", "subsea", "experimental", "europe", "research" ], "last_updated": "2026-06-24" } ]