PUBLIC LANDS WATCH

Artificial Intelligence as a Driver of Public-Lands Threats

Nationwide — federal sites named to date are Idaho National Laboratory, Oak Ridge Reservation, Paducah Gaseous Diffusion Plant, and Savannah River Site · Cross-cutting policy threat (not a single site) — federal energy, leasing, minerals, water, and permitting policy tied to AI infrastructure · managed by Department of Energy / Department of the Interior / Department of Defense / U.S. Customs and Border Protection

Status: Active — orders in force, federal sites selected, no BLM public-land site identified Threat: Agency capacity — federal policy linking AI infrastructure to public-land use
Scope
System-wide policy file — not a single location
Coordinates
38.8899, -77.0091
Page last verified
2026-08-17
Site maintainer of record
Alicen M.
Last edit
2026-08-17 by Alicen M.
File
/ai-data-centers-public-lands

01 — Historical & Cultural Record

This file tracks a policy chain rather than a landscape. The instruments below are the ones that connect artificial-intelligence infrastructure to decisions about federal land, listed with the dates and, where captured, the Federal Register citations.

A citation conflict worth recording. Several law-firm analyses cite the data-center permitting order variously as Executive Order 14318, 14148, or 14149. The White House text is authoritative and the most consistent citation across sources is 14318, which this page uses. The discrepancy is noted rather than silently resolved.23

02 — Current Threat

Angle 1 — Energy and data centers.

The Department of Energy's 2024 U.S. Data Center Energy Usage Report, prepared by Lawrence Berkeley National Laboratory (Shehabi et al., December 2024, LBNL-200163), found U.S. data centers consumed 176 TWh in 2023 — 4.4% of total U.S. electricity consumption — and projected roughly 325 to 580 TWh by 2028, or 6.7% to 12.0% of national consumption, a compound annual growth path of 13% to 27%.7

The Energy Information Administration's Short-Term Energy Outlook puts total U.S. electricity use rising from a record 4,195 billion kWh in 2025 to 4,271 billion kWh in 2026 and 4,397 billion kWh in 2027; its Annual Energy Outlook 2026, published in April, shows the server share of commercial-sector electricity climbing over the long term. These EIA figures are recorded here as reported secondhand; the STEO release itself has not been retrieved for this page.17

Executive Order 14318 (July 23, 2025) defines a "Data Center Project" as "a facility that requires greater than 100 megawatts (MW) of new load dedicated to AI inference, training, simulation, or synthetic data generation." Section 9 directs the Departments of the Interior and Energy to identify appropriate federal sites, and directs the Secretary of Defense to identify suitable sites on military installations. Section 4 revokes Executive Order 14141 of January 14, 2025.23

Federal sites selected. On July 24, 2025 the Department of Energy named four federal sites for AI data centers with co-located generation: Idaho National Laboratory, the Oak Ridge Reservation, the Paducah Gaseous Diffusion Plant, and the Savannah River Site. Energy Secretary Chris Wright stated: "By leveraging DOE land assets for the deployment of AI and energy infrastructure, we are taking a bold step to accelerate the next Manhattan Project — ensuring U.S. AI and energy leadership." The announcement records that the selections followed an April 2025 request for information to which DOE "received enormous interest"; it does not state a number of candidate sites, and the figure of 16 candidates carried in earlier research for this page could not be confirmed against it.43

Transmission. DOE's draft 2026 National Transmission Needs Study states that rapid load growth, led by AI data centers, has become the principal driver of new transmission investment. It cites LBNL, EPRI, and McKinsey projections of more than 400 TWh of U.S. data-center demand by 2030, and NERC forecasts of total U.S. consumption rising from 4,281 TWh in 2024 to 5,353 TWh in 2034.18 LBNL reports interconnection queues exceeding 1,500 GW as of 2025.7

The gap that matters most for this registry. Per a Congressional Research Service product and law-firm analyses, the Department of Energy and the Air Force are accepting project proposals, but the Department of the Interior has not publicly identified any BLM public-land site for data-center development under Executive Order 14318. No BLM right-of-way application and no Federal Register notice for a data center, or for associated generation or transmission, on BLM public land was located in this research pass — a finding consistent with that gap.319 Specific FERC large-load docket numbers were likewise not captured with primary sources.

Angle 6 — AI and the permitting process itself. Executive Order 14318 lowers the review burden in two separate places. Section 5, "Efficient Environmental Reviews," directs at 5(b) that "the Council on Environmental Quality shall coordinate with relevant agencies on the establishment of new categorical exclusions to cover actions related to Qualifying Projects that normally do not have a significant effect on the human environment" — that is, creating new exclusions rather than applying existing ones. Section 5(c) provides that federal financial assistance representing less than 50% of project costs is not a "major Federal action" under NEPA. Separately, section 8, "Biological and Water Permitting Efficiencies," directs the Secretaries of the Interior and Commerce to use programmatic consultation for common construction activities on Qualifying Projects "that will occur over the next 10 years." The broader NEPA-streamlining posture is set by Executive Order 14154 and the associated rollback of Council on Environmental Quality NEPA regulations.23

Two threads on this angle are open and explicitly unfilled: named DOE or Pacific Northwest National Laboratory NEPA-AI document-analysis tools, and the Department of the Interior AI use-case inventory required by federal AI policy, were not located with primary sources. No GAO or Inspector General report specifically flagging AI-accelerated environmental-review concerns was located either.

03 — Documented Impact: Environmental

Angle 2 — the fossil-fuel and leasing linkage.

Executive Order 14261 is the clearest official document tying AI data-center demand to federal fossil-fuel leasing. Section 1 states: "Our Nation's beautiful clean coal resources will be critical to meeting the rise in electricity demand due to the resurgence of domestic manufacturing and the construction of artificial intelligence data processing centers." Section 5, "Lifting Barriers to Coal Mining on Federal Lands," directs Interior and Agriculture to prioritize and expedite coal leasing on federal lands, and section 5(b) directs Interior to "acknowledge the end of the Jewell Moratorium" by publishing a Federal Register notice terminating the environmental impact statement that maintained it. Section 10 is titled "Powering Artificial Intelligence Data Centers" and directs Interior, Commerce, and Energy to identify regions where coal infrastructure can support AI data centers and to assess expanding coal-based power for AI and high-performance computing.1

The Department of Energy's Resource Adequacy Report of July 7, 2025, issued pursuant to Executive Order 14262, estimates roughly 100 GW of new peak supply needed by 2030, of which about 50 GW is directly attributable to data centers, and frames preventing coal and gas retirements as necessary to "win the artificial intelligence (AI) arms race."5

On September 29, 2025 Interior opened 13.1 million acres to coal leasing and cut the federal coal royalty rate to 7%, citing Executive Order 14261 as authority. This page keeps a precision note the record requires: the Interior and BLM release announcing that action does not itself mention AI or data centers. The AI rationale lives in the executive order the release invokes. See /sites/federal-coal-leasing/.61

A second precision note applies to the AI Action Plan. "Winning the Race: America's AI Action Plan" (July 2025) ties AI to dispatchable power generation — coal, gas, geothermal, nuclear — and to grid reliability, but does not contain an explicit instruction to expand oil, gas, or coal leasing on federal land for AI. This page does not treat it as though it does.

Separately, the Department of Energy has issued more than 40 emergency orders under section 202(c) of the Federal Power Act since May 2025, keeping upward of 4.5 GW of coal generation online. The orders themselves are grounded in Federal Power Act reliability emergencies; the AI framing appears in DOE's Resource Adequacy Report and in Secretary statements rather than verbatim in every order.245

Angle 4 — water.

The same DOE/LBNL 2024 report records that U.S. data centers directly consumed approximately 17.4 billion gallons of water in 2023, up from roughly 5.6 billion in 2014, and projects 38 to 73 billion gallons by 2028, with hyperscale facilities accounting for about half. LBNL also finds indirect water use — power-plant cooling upstream of the data center — runs roughly 12 times direct cooling use.2021

Facility-level figures on the record:

A Congressional Research Service product dated July 31, 2026 puts data centers at roughly 2% of total U.S. water consumption, states that a single 100-MW facility may directly consume water equivalent to about 2,600 households per day, and records that the federal government has conducted no systematic assessment and maintains no comprehensive federal data collection on data-center water use.12 A separate CRS report gives the International Energy Agency's figure of roughly 2 million liters per day for a 100-MW facility.13

Arizona and California require water-efficiency evaluation — recirculating or dry cooling — in power-plant siting, which is the primary regulatory hook near many Western public lands.21

Conflict flag on the water figures. Direct 2023 U.S. data-center water consumption is reported as 17.4 billion gallons by LBNL and EPA and about 17.5 billion gallons by IEEE Spectrum — effectively the same figure.2023 A separate Nature-based figure of 449 million gallons per day, or about 163.7 billion gallons per year for 2021, is far higher and measures on a different basis. This page treats LBNL's 17.4 billion gallons (direct, 2023) as the primary figure and records the others rather than averaging them. Water-rights applications specifically adjacent to named public lands, and tribal water disputes tied to data-center siting, were not located with primary sources in this pass.

03 — Documented Impact: Cultural & Archaeological

Angle 5 — surveillance and border technology on public and tribal lands.

Anduril Industries won a $363 million contract from U.S. Customs and Border Protection for more than 200 Extended Range Sentry Towers: 80-foot AI-enabled autonomous surveillance towers carrying electro-optical, thermal, and radar sensors feeding the company's Lattice AI software. The one-year contract was awarded in December 2025 and announced June 13, 2026. Anduril had delivered more than 40 of the towers and was building more than 15 a month. CBP already operates over 350 Standard Range Sentry towers.162514

The Extended Range towers detect, classify, and track objects beyond 5 miles, and up to roughly 7.5 miles (12 km) with an operator. Anduril estimates its autonomous surveillance towers provide about 30% coverage of the U.S. southern land border — a company estimate, reaffirmed with the June 2026 announcement, not an agency finding.1626

These towers are deployed across the southwest border including the Big Bend region, where CBP's "smart wall" approach substitutes cameras, sensors, and patrol roads for a physical wall inside a national park. See /sites/big-bend/. A June 2026 GAO bid-protest ruling addressed the IIRIRA section 102(c) waivers that suspended competition rules for border task orders.28

Conflict flag on the contract value. The award is reported as $363 million by most sources and as at least $367 million by Bloomberg Government, which describes it as an SBIR Phase III order under the Autonomous Surveillance Towers IDIQ with $42 million slid against it. The difference appears to reflect base-award versus task-order accounting. This page uses $363 million as the announced figure. USASpending and FPDS award records that would settle it were not retrieved.1627

Documented gap. Tower deployments specifically inside Organ Pipe Cactus National Monument, Cabeza Prieta, Buenos Aires, Lower Rio Grande Valley, and San Bernardino National Wildlife Refuges were not confirmed with primary sources, and no official statement of the Tohono O'odham Nation was captured in this pass. Those are open research questions, not findings that no deployments or statements exist.

03 — Documented Impact: Economic & Fiscal

Angle 3 — critical minerals and semiconductors.

The USGS Final 2025 List of Critical Minerals, published in the Federal Register on November 7, 2025 (doc. 2025-19813), names 60 minerals, adding ten to the 2022 list: boron, copper, lead, metallurgical coal, phosphate, potash, rhenium, silicon, silver, and uranium. Copper's inclusion was long sought, and uranium's re-designation bears directly on the mining and milling record tracked at /sites/bears-ears/ and /sites/grand-canyon-pinyon-plain/.89

Minerals on the list relevant to semiconductors and data-center hardware include copper, gallium, germanium, silicon, cobalt, and the rare earths. Gallium and germanium are among the highest supply-risk commodities under USGS's probability-weighted methodology.10

The policy drivers behind the list: Executive Order 13817 (2017) started it; Executive Order 14154 (January 20, 2025) directed USGS to consider adding uranium; Executive Order 14261 (April 8, 2025) directed consideration of metallurgical coal; an executive order of April 15, 2025 ordered a Section 232 national-security investigation into imported processed critical minerals; and Executive Order 14285 (April 2025) addresses seabed minerals.82911

Copper supply context. USGS reports U.S. copper mines produced roughly 1 million tonnes in 2025, down 5% year over year, and the International Energy Agency warns of a potential 25% global primary-copper deficit by 2035 — the backdrop against which pressure on domestic copper mines is building.930

Documented gap, stated plainly. Specific FAST-41 Permitting Dashboard covered-project entries for named mines on public land, and Department of Defense or Defense Production Act Title III award records tied explicitly to AI or chip demand, were not located in this research pass. Resolution Copper's current status, and any 2026 official filing linking it to AI or chip demand, were not confirmed with a primary source. Alongside the permitting-tools thread in the Current Threat section, this is the thinnest angle on this page and the one most in need of follow-up.

03 — Documented Impact: Legal Status

Executive Order 14318 is the operative instrument for environmental review. Section 5, "Efficient Environmental Reviews," directs the Council on Environmental Quality to coordinate on establishing new NEPA categorical exclusions for Qualifying Projects. Section 8, "Biological and Water Permitting Efficiencies," directs Interior and Commerce to use programmatic Endangered Species Act consultation for common construction activities occurring "over the next 10 years."2

No litigation over the data-center permitting order, the federal site selections, or the AI provisions of Executive Order 14261 has been identified in current sourcing as of August 17, 2026. This is an open research question rather than a finding that no challenge exists.

The related litigation this registry does track sits on the individual case-file pages: the two Big Bend waiver suits at /sites/big-bend/, and the Arctic refuge leasing challenge at /sites/arctic-refuge/.

03 — Documented Impact: Human Health

Not yet documented — no health-outcomes study specific to AI-driven federal energy, minerals, or water policy was identified in current sourcing. Health findings tied to the individual extraction sites this policy chain touches are documented on those pages, particularly /sites/bears-ears/ and /sites/grand-canyon-pinyon-plain/.

03 — Tribal / Public Consultation Status

Not yet documented. No tribal or public consultation record specific to the federal site selections of July 24, 2025, to Executive Order 14318's implementation, or to the AI provisions of Executive Order 14261 was identified in current sourcing.

Consultation records for the individual actions taken under these orders are documented on the relevant case-file pages, where the record is generally one of compressed or absent process — see /sites/chaco-culture/, /sites/bears-ears/, and /sites/tongass-roadless-rule/.

04 — Named Companies & Responsible Parties

05 — Recorded Opposition & Support

06 — Timeline of Developments

07 — Sources

Primary

Secondary

Corrections Log

2026-08-17 — corrections from a primary-source verification pass. Four claims carried in the research this page was built from did not survive checking against the White House and DOE texts, and have been changed:

Also corrected: Executive Order 14261 section 5(b) directs Interior to "acknowledge the end of" the Jewell Moratorium by publishing a notice terminating the supporting environmental impact statement, not to end the moratorium itself; and the Roadless Rule notice of intent cites four executive orders, not three (14192 was missing).

Verified unchanged: Executive Order 14261 section 1's quoted sentence, the section 5 title, and the section 10 title "Powering Artificial Intelligence Data Centers"; the four DOE sites and the July 24, 2025 date; the 60-mineral 2025 critical minerals list and its ten additions.

Take Action

No open federal comment period specific to this policy chain is on record as of last verification. Comment periods for the individual actions taken under these orders are listed on the relevant case-file pages.

Sign / comment on the public record for Artificial Intelligence as a Driver of Public-Lands Threats →