{ "schemaVersion": "1.0.0", "identity": { "dossierId": "DOS-1", "translationKey": "infraestrutura-datacenters-overview", "title": "Data centers — overview of infrastructure, expansion, and limits", "type": "Overview", "locale": "en", "sourceLocale": "pt-br", "localizationStatus": "translated", "slug": "data-centers-infrastructure-expansion-limits", "intent": "explain", "audience": "Readers, researchers, investors, operators, and policymakers.", "scope": "How digital demand becomes physical infrastructure; which resources and bottlenecks constrain expansion; and how costs, impacts, and power are distributed.", "exclusions": [ "Technical, environmental, fiscal, or legal details remain in the specialized Dossiers.", "The Overview does not reproduce the complete evidence bases of the four specialized Dossiers." ], "thesis": "Data centers are not abstract digital infrastructure. They are cyber-physical systems whose expansion depends on the convergence of compute, power, cooling, connectivity, capital, equipment, territory, and authorization." }, "relations": { "specializedDossiers": [ { "dossierId": "DOS-6", "title": "Data centers — how they work and why they are expanding", "slug": "datacenters-sistema-fisico-demanda" }, { "dossierId": "DOS-2", "title": "Data centers — energy and environmental impacts", "slug": "datacenters-energia-impactos-ambientais" }, { "dossierId": "DOS-3", "title": "Data centers — political economy, investment, and local impacts", "slug": "datacenters-economia-politica-impactos-locais" }, { "dossierId": "DOS-4", "title": "Data centers — geography, geopolitics, and sovereignty", "slug": "datacenters-geografia-geopolitica-soberania" } ] }, "executiveSummary": [ "Data centers are physical systems that combine servers, accelerators, memory, storage, and networks with buildings, power, cooling, connectivity, capital, and authorization. The sector's expansion cannot be explained by a single component. AI accelerates part of the demand; power has become a significant constraint; chips, networks, electrical equipment, territory, and regulation remain interdependent.", "Announced, contracted, installed, energized, usable, and actually consumed capacity are distinct states.", "Measurement gaps prevent rankings and false precision, but they do not prevent explanations of known mechanisms." ], "systemMap": "demand for digital services\n ↓\nworkloads + delivery model\n ↓\nservers, accelerators, memory, storage, and networking\n ↓\nracks, critical power, cooling, and building\n ↓\ngrid connection + fiber + cables + territory\n ↓\ncontracts, financing, incentives, and regulation\n ↓\ninstalled → energized → usable → actually consumed capacity\n ↓\neconomic, environmental, territorial, and strategic effects", "data": [ { "id": "DC-OV-D01", "title": "A data center combines IT and supporting infrastructure", "claim": "A data center combines servers, storage, and networking with power supply, conversion, redundancy, cooling, security, and a building.", "types": [ "fact" ], "status": "verified", "limitations": [ "Implementations vary by facility type and availability requirements." ], "confidence": "high", "lastReviewed": "2026-07-25", "references": [ "src-epa-energy-star-datacenter-faq-pue", "src-lbnl-us-data-center-energy-2024" ] }, { "id": "DC-OV-D02", "title": "Cloud is not a workload", "claim": "Cloud is a delivery and operating model; AI, video, databases, web services, analytics, HPC, and storage are applications or workload classes that can share the same infrastructure.", "types": [ "fact", "taxonomy" ], "status": "verified", "limitations": [ "Application categories can overlap in their use of resources, making mutually exclusive shares difficult to determine." ], "confidence": "high", "references": [ "src-nist-sp800-145-cloud-definition" ] }, { "id": "DC-OV-D03", "title": "Global projections are specific to an edition and scenario", "claim": "The IEA estimated about 415 TWh for data centers in 2024 in the 2025 Energy and AI edition; a later update used approximately 485 TWh in 2025 and a central scenario close to 950 TWh in 2030.", "types": [ "estimate", "prediction" ], "status": "contextual", "period": "Base years 2024 and 2025; central scenario for 2030.", "limitations": [ "The values belong to different editions, base years, and assumptions; they do not form a single range." ], "confidence": "high for attribution; medium for projection", "misinterpretationRisk": "Do not present 2030 as a future fact.", "references": [ "src-iea-energy-and-ai-2025", "src-iea-key-questions-energy-ai-2026" ] }, { "id": "DC-OV-D04", "title": "Concentration makes local impact disproportionate", "claim": "Data centers accounted for 23% of Ireland's measured electricity consumption in 2025, compared with 5% in 2015.", "types": [ "fact" ], "status": "verified", "measurementBasis": "Metered electricity consumption.", "period": "2015–2025", "limitations": [ "This is not a share of all final energy and is not a global measure." ], "confidence": "high", "references": [ "src-cso-ireland-datacenter-electricity-2025" ] }, { "id": "DC-OV-D05", "title": "An AI-dense configuration does not represent a typical rack", "claim": "NVIDIA documentation reports approximately 120 kW for an NVL72 configuration; in the 2025 Uptime survey, 82% of 709 responding operators did not have a rack above 30 kW as the highest-density rack in the facility.", "types": [ "comparison" ], "status": "contextual", "measurementBasis": "Specification for one NVL72 configuration compared with a survey of 709 operators.", "period": "Uptime 2025 survey.", "limitations": [ "A product specification and an operator survey cover different populations; the survey is not a distribution of all racks." ], "confidence": "high for the sources; low for global generalization", "references": [ "src-nvidia-dgx-gb-rack-hardware-guide", "src-uptime-global-datacenter-survey-2025" ] }, { "id": "DC-OV-D06", "title": "TDP and benchmarks do not measure annual site consumption", "claim": "TDP is a component property; MLPerf Power measures a system's AC power during a defined benchmark; neither measure, in isolation, provides a data center's annual consumption.", "types": [ "comparison" ], "status": "verified", "measurementBasis": "Component TDP and system AC power during a defined benchmark.", "limitations": [], "confidence": "high", "references": [ "src-nvidia-h100-specifications", "src-mlcommons-mlperf-inference", "src-mlcommons-mlperf-power" ] }, { "id": "DC-OV-D07", "title": "Water use depends on boundaries and conditions", "claim": "An LBNL study found a variation of more than four orders of magnitude in water consumption per workload among the evaluated combinations of equipment, utilization, cooling, climate, power grid, and hardware refresh.", "types": [ "estimate", "comparison" ], "status": "contextual", "measurementBasis": "Water consumption per workload among the evaluated combinations.", "limitations": [ "This does not represent the observed distribution of global sites and does not justify a universal number 'per query.'" ], "confidence": "medium-high", "references": [ "src-lbnl-water-use-workloads-2025" ] }, { "id": "DC-OV-D08", "title": "A renewable contract does not prove hourly physical supply", "claim": "Market-based Scope 2 records contractual instruments; on its own, it does not demonstrate that consumption was supplied by carbon-free generation in every hour and location.", "types": [ "comparison" ], "status": "verified", "limitations": [ "Distinguish the current rule from future proposals for hourly matching and deliverability." ], "confidence": "high", "references": [ "src-ghg-protocol-scope2-guidance-2015" ] }, { "id": "DC-OV-D09", "title": "In the Virginia case, modeled economic impact is predominantly construction", "claim": "The Virginia model estimated 74,000 jobs associated with the sector: 59,000 tied to construction and 15,000 to operations, including direct, indirect, and induced effects.", "types": [ "estimate" ], "status": "contextual", "measurementBasis": "Modeled jobs, including direct, indirect, and induced effects.", "limitations": [ "This is not equivalent to permanent employees of operators and does not represent other jurisdictions." ], "confidence": "high for attribution", "references": [ "src-jlarc-data-centers-virginia-2024" ] }, { "id": "DC-OV-D10", "title": "Use of an incentive does not prove additionality", "claim": "The Washington audit found fewer than ten beneficiaries, only refurbishments, no new construction under the preference, and an estimated US$42.4 million in savings between 2023–2026.", "types": [ "fact", "estimate" ], "status": "verified Case Study", "period": "2023–2026", "limitations": [ "This does not measure every incremental investment and does not support national generalization." ], "confidence": "high", "references": [ "src-wa-jlarc-urban-data-centers-tax-preference-2026" ] }, { "id": "DC-OV-D11", "title": "Regional electricity consumption is not a share of compute", "claim": "For 2024, the IEA attributed 45%, 25%, and 15% of estimated data-center electricity consumption to the United States, China, and Europe.", "types": [ "estimate" ], "status": "contextual", "measurementBasis": "Share of estimated data-center electricity consumption.", "period": "2024", "limitations": [ "This does not measure GPUs, FLOPs, servers, or usable capacity." ], "confidence": "high for the published estimate", "references": [ "src-iea-energy-and-ai-2025" ] }, { "id": "DC-OV-D12", "title": "Cables are critical, but cable count does not measure resilience", "claim": "The ITU estimates that submarine cables carry more than 99% of international data traffic; faults, repairs, capacity, route diversity, and unavailability are distinct metrics.", "types": [ "estimate", "comparison" ], "status": "contextual", "measurementBasis": "Estimated share of international data traffic.", "limitations": [], "confidence": "medium-high", "references": [ "src-itu-submarine-cable-resilience-2026" ] }, { "id": "DC-OV-D13", "title": "Location is not comprehensive sovereignty", "claim": "Keeping data and data centers within a territory does not, on its own, determine effective jurisdiction, operational control, keys, portability, technological substitutability, or continuity.", "types": [ "comparison" ], "status": "verified as framework", "limitations": [], "confidence": "high", "references": [ "src-oecd-data-localisation-2020" ] } ], "info": [ { "id": "DC-OV-I01", "title": "Expansion is the coordination of dependencies", "interpretation": "Chips without power create unproductive inventory; power without hardware, connectivity, or demand creates idle capacity; neither becomes operational without permits and contracts.", "derivedFrom": [ "DC-OV-D01", "DC-OV-D05", "DC-OV-D06" ] }, { "id": "DC-OV-I02", "title": "The bottleneck moves through the system", "interpretation": "The constraint can move among chips, HBM, packaging, networking, electrical equipment, grid connection, cooling, construction, and commissioning.", "derivedFrom": [ "DC-OV-D01", "DC-OV-D05" ], "provenanceNote": "The canonical derivation also points to the specialized Dossiers on Operations and Energy, without identifying additional claims in the Overview." }, { "id": "DC-OV-I03", "title": "Unit efficiency and aggregate consumption can diverge", "interpretation": "Lower energy use per task does not determine total consumption; volume, price, utilization, and new uses also change.", "derivedFrom": [ "DC-OV-D03", "DC-OV-D06" ] }, { "id": "DC-OV-I04", "title": "Small global impacts can be large locally", "interpretation": "A global share of electricity or water does not describe pressure on a grid node, watershed, or municipality.", "derivedFrom": [ "DC-OV-D04", "DC-OV-D07", "DC-OV-D11" ] }, { "id": "DC-OV-I05", "title": "Benefits and costs appear in different institutions", "interpretation": "A state tax expenditure, municipal revenue, utility investment, electricity rates, and temporary employment can coexist; they should not be added together as a net benefit without a counterfactual.", "derivedFrom": [ "DC-OV-D09", "DC-OV-D10" ] }, { "id": "DC-OV-I06", "title": "Location is simultaneously an advantage and a dependency", "interpretation": "Power, fiber, and regulation attract capacity, but chips, cloud, software, cables, and jurisdiction preserve external dependencies.", "derivedFrom": [ "DC-OV-D11", "DC-OV-D12", "DC-OV-D13" ] }, { "id": "DC-OV-I07", "title": "A projection or incentive does not demonstrate delivered capacity", "interpretation": "Demand projections and incentive programs do not, on their own, demonstrate operational capacity; deployment still depends on financing, equipment, grid connection, permits, construction, and operations.", "derivedFrom": [ "DC-OV-D03", "DC-OV-D10" ] } ], "uncertainties": [ { "id": "DC-OV-UNC-01", "title": "Global share by workload", "status": "not publicly measured", "known": "Adoption and presence are not equivalent to energy, servers, or compute share.", "treatment": "Do not publish an artificial global breakdown.", "confidence": "high confidence in the gap", "lastReviewed": "2026-07-25" }, { "id": "DC-OV-UNC-02", "title": "Installed versus usable capacity", "status": "not publicly measured", "known": "Power, cooling, space, networks, and reliability can produce stranded capacity.", "treatment": "Explain the mechanisms without a global percentage.", "lastReviewed": "2026-07-25" }, { "id": "DC-OV-UNC-03", "title": "Comparable environmental impact by site", "status": "incompatible methods", "known": "PUE, WUE, water, emissions, and embodied carbon depend on boundary, climate, and operations.", "treatment": "Preserve the site, period, denominator, and method.", "lastReviewed": "2026-07-25" }, { "id": "DC-OV-UNC-04", "title": "Net fiscal return", "status": "not publicly measured", "known": "Gross revenue and tax expenditures exist at different levels of government.", "treatment": "Do not produce an incentive ranking.", "lastReviewed": "2026-07-25" }, { "id": "DC-OV-UNC-05", "title": "Global compute and sovereignty", "status": "imperfect proxy", "known": "Electricity, racks, facilities, and location measure partial dimensions.", "treatment": "Use a multidimensional framework, not a single index.", "lastReviewed": "2026-07-25" } ], "methodology": { "data": "DATA records verifiable claims with explicit sources, measurement context, limitations, and confidence.", "info": "INFO records synthesis or interpretation and always declares the IDs of the claims from which it is derived.", "uncertainty": "An Uncertainty records a gap as an editorial result; it remains explicit and is not filled with a fragile estimate.", "updateRule": "A review may correct or update the projection only from the canonical Dossier with editorial authorization; material changes increment the public version and are recorded in the corrections history." }, "version": { "publicVersion": "1.0.0", "lastVerified": "2026-07-29", "staleAfter": "2026-08-24", "corrections": [] }, "references": [ { "id": "src-epa-energy-star-datacenter-faq-pue", "title": "U.S. EPA ENERGY STAR — Data center rating FAQ and PUE definition", "publisher": "U.S. EPA", "url": "https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P100TSZE.TXT", "accessedAt": "2026-07-24" }, { "id": "src-lbnl-us-data-center-energy-2024", "title": "LBNL — 2024 United States Data Center Energy Usage Report", "publisher": "Lawrence Berkeley National Laboratory", "url": "https://eta-publications.lbl.gov/publications/2024-lbnl-data-center-energy-usage-report", "accessedAt": "2026-07-24" }, { "id": "src-nist-sp800-145-cloud-definition", "title": "NIST — The Definition of Cloud Computing (SP 800-145)", "publisher": "NIST", "url": "https://csrc.nist.gov/pubs/sp/800/145/final", "accessedAt": "2026-07-24" }, { "id": "src-iea-energy-and-ai-2025", "title": "IEA — Energy and AI (2025)", "publisher": "IEA", "url": "https://www.iea.org/reports/energy-and-ai", "accessedAt": "2026-07-22" }, { "id": "src-iea-key-questions-energy-ai-2026", "title": "IEA — Key Questions on Energy and AI (2026)", "publisher": "IEA", "url": "https://www.iea.org/reports/key-questions-on-energy-and-ai", "accessedAt": "2026-07-24" }, { "id": "src-cso-ireland-datacenter-electricity-2025", "title": "CSO Ireland — Data Centres Metered Electricity Consumption 2025", "publisher": "Central Statistics Office Ireland", "url": "https://www.cso.ie/en/releasesandpublications/ep/p-dcmec/datacentresmeteredelectricityconsumption2025/", "accessedAt": "2026-07-24" }, { "id": "src-nvidia-dgx-gb-rack-hardware-guide", "title": "NVIDIA — DGX GB Rack Scale Systems Hardware Guide", "publisher": "NVIDIA", "url": "https://docs.nvidia.com/dgx/dgxgb200-user-guide/hardware.html", "accessedAt": "2026-07-24" }, { "id": "src-uptime-global-datacenter-survey-2025", "title": "Uptime Institute — 2025 Global Data Center Survey", "publisher": "Uptime Institute", "url": "https://uptimeinstitute.com/about-ui/press-releases/uptimes-15th-annual-global-data-center-survey-results-shows-both-commitment-and-hesitancy", "accessedAt": "2026-07-24" }, { "id": "src-nvidia-h100-specifications", "title": "NVIDIA — H100 Tensor Core GPU specifications", "publisher": "NVIDIA", "url": "https://www.nvidia.com/en-us/data-center/h100/", "accessedAt": "2026-07-22" }, { "id": "src-mlcommons-mlperf-inference", "title": "MLCommons — MLPerf Inference benchmark documentation", "publisher": "MLCommons", "url": "https://docs.mlcommons.org/inference/", "accessedAt": "2026-07-24" }, { "id": "src-mlcommons-mlperf-power", "title": "MLCommons — MLPerf Inference power measurement methodology", "publisher": "MLCommons", "url": "https://docs.mlcommons.org/inference/power/", "accessedAt": "2026-07-24" }, { "id": "src-lbnl-water-use-workloads-2025", "title": "LBNL — The water use of data center workloads (2025)", "publisher": "Lawrence Berkeley National Laboratory", "url": "https://eta-publications.lbl.gov/publications/water-use-data-center-workloads", "accessedAt": "2026-07-24" }, { "id": "src-ghg-protocol-scope2-guidance-2015", "title": "GHG Protocol — Scope 2 Guidance", "publisher": "GHG Protocol / WRI", "url": "https://ghgprotocol.org/scope-2-guidance", "accessedAt": "2026-07-24" }, { "id": "src-jlarc-data-centers-virginia-2024", "title": "JLARC — Data Centers in Virginia (2024)", "publisher": "Virginia JLARC", "url": "https://jlarc.virginia.gov/landing-2024-data-centers-in-virginia.asp", "accessedAt": "2026-07-24" }, { "id": "src-wa-jlarc-urban-data-centers-tax-preference-2026", "title": "Washington JLARC — Data Centers in Urban Counties: 2026 Tax Preference Review", "publisher": "Washington JLARC", "url": "https://leg.wa.gov/jlarc/taxReports/2026/data/1/report.html", "accessedAt": "2026-07-24" }, { "id": "src-itu-submarine-cable-resilience-2026", "title": "ITU — International Advisory Body on Submarine Cable Resilience: Working Groups Report 2026", "publisher": "ITU", "url": "https://www.itu.int/hub/publication/s-iab-wg-2026/", "accessedAt": "2026-07-24" }, { "id": "src-oecd-data-localisation-2020", "title": "OECD — Data localisation trends and challenges", "publisher": "OECD", "url": "https://www.oecd.org/content/dam/oecd/en/publications/reports/2020/12/data-localisation-trends-and-challenges_d775fe8a/7fbaed62-en.pdf", "accessedAt": "2026-07-24" } ] }