AI DATA CENTERS, FEDERAL CONTROL, AND THE NEW COMPUTE-STATE POWER FIGHT
Evidence-Driven Research Dossier
Prepared as a multidisciplinary PhD-level research brief
Date: June 25, 2026
The issue is not just who owns the servers. The issue is who owns the compute, who controls the power, who writes the rules, who gets the data, and who decides what reality the machine is allowed to show us.
Table of Contents
1. Executive Summary
2. What Data Centers Actually Are
3. Why AI Changed the Data Center Arena
4. Major Companies and Government Players
5. Bernie Sanders 50% Ownership Claim
6. Trump / Data Centers as Military Installations Claim
7. Energy, Water, and Grid Pressure
8. Legal and Constitutional Issues
9. National Security and Military Angle
10. Surveillance, Censorship, and Civil Liberties
11. Follow the Money
12. Local Community Impact
13. Data Centers as the New Military-Industrial-Tech Complex
14. Claim-by-Claim Fact Check
15. Red Flags and Green Flags
16. Source Ranking
17. Final Bottom-Line Conclusion
Appendix A. Key Source Notes
1. Executive Summary
Bottom line: AI data centers are becoming strategic infrastructure because modern AI is no longer just software. It is software plus chips, electricity, water, land, fiber, security, capital, and political permission. Whoever controls the compute stack controls how fast frontier AI advances, who gets access, who gets priced out, what runs on the machines, and how closely government can attach itself to private AI infrastructure.
The two headline claims break this way:
| Claim | Verdict | What the evidence shows |
| Bernie Sanders wants the U.S. government/public to own 50% of major AI companies. | Verified as proposed legislation, not enacted law. | Sanders introduced/announced the American A.I. Sovereign Wealth Fund Act, structured as a one-time 50% tax paid in equity stock by large/systemically important AI companies. The stock would go into a federal sovereign wealth fund managed by an Independent Commission for Democratic AI. It includes AI companies tied to AI data centers, AI computing infrastructure, AI services, and advanced robotics. |
| Trump declared data centers military installations. | Misleading / not verified as a literal legal declaration. | Trump’s July 2025 executive order treats AI data centers and supporting power infrastructure as important to national security and accelerates federal permitting. Separately, the Army and Air Force have pursued leasing land on military bases for private/commercial AI data centers. That is military-adjacent infrastructure, not proof that all data centers have been legally declared military installations. |
The real story is bigger than either politician. AI infrastructure is being pulled into a new triangle: Big Tech, federal power, and energy infrastructure. Sanders’ approach is public ownership/control through forced equity taxation. Trump’s approach, based on the documents found, is accelerated buildout, federal land access, national-security framing, and military-base leasing - not nationalization of data centers. Both tracks treat AI compute as too important to leave in the normal commercial lane.
2. What Data Centers Actually Are
A normal data center is a facility that houses servers, networking gear, storage systems, cooling equipment, backup power, security systems, and fiber connections. A modern AI data center is that idea on steroids.
Traditional server farm: runs websites, email, business software, storage, video, cloud services, and ordinary enterprise workloads.
AI data center: runs GPU or accelerator-heavy workloads for training and inference. That means huge power density, specialized cooling, high-speed networking, and massive clusters of chips working together.
| Type | Plain-English meaning |
| Hyperscale | Giant facilities built for companies like Microsoft, Amazon, Google, Meta, Oracle, and major AI labs. |
| Colocation | A company owns the building and rents secure server space, power, and cooling to other customers. Equinix and Digital Realty are major examples. |
| Edge data center | Smaller facility closer to users for lower latency. Think local or regional compute. |
| AI training data center | Built to train large models. Heavy GPU use, huge power draw, intense networking demands. |
| AI inference data center | Runs already-trained models for users. This becomes massive when millions of people use AI every day. |
| Cloud data center | Provides compute, storage, and networking over the internet through platforms like AWS, Azure, Google Cloud, and Oracle Cloud. |
The Trump executive order defines a covered “Data Center Project” as a facility requiring more than 100 MW of new load dedicated to AI inference, training, simulation, or synthetic data generation. That threshold matters: 100 MW is not a small server room. It is an industrial-scale electric load.
3. Why AI Changed the Data Center Arena
AI made compute strategic. Old cloud computing was mostly about storage, websites, apps, databases, and enterprise software. AI infrastructure is different because frontier AI depends on concentrated compute: GPUs, TPUs, networking, cooling, and electricity. The bottleneck is no longer just programmers. It is chips, power, land, transformers, substations, water, and capital.
Chips: NVIDIA GPUs dominate frontier AI training and inference. AMD, Intel, Google TPUs, Amazon Trainium, Microsoft Maia, and other custom accelerators are attempts to reduce dependence on NVIDIA.
Clusters: Thousands to hundreds of thousands of accelerators must be networked together.
Power: AI clusters consume massive electricity continuously.
Cooling: Dense GPU racks generate heat that traditional air cooling cannot always handle efficiently.
Data: Training and fine-tuning require enormous datasets.
Cloud control: The companies that own cloud infrastructure can control access, pricing, deployment, moderation, surveillance compliance, and model availability.
Export controls: The U.S. has restricted advanced AI chips to China and other adversarial destinations because compute is now treated as dual-use strategic capability.
That is why AI compute is starting to look like earlier strategic assets: oil, steel, railroads, telecom, semiconductors, nuclear material, and satellite infrastructure. Not because servers are magic, but because modern power flows through them.
4. Major Companies and Government Players
| Player | Role | Incentive | What they control |
| OpenAI | Frontier AI lab | Model dominance, enterprise revenue, cloud scale | Models, APIs, partnerships, safety rules |
| Anthropic | Frontier AI lab | Enterprise AI, safety-branded model deployment | Claude models, enterprise systems |
| xAI | Frontier AI lab | Model competition, social/data integration | Grok ecosystem, AI compute demand |
| Google / DeepMind / Google Cloud | AI lab + cloud + chips | Full-stack AI dominance | TPUs, data, cloud, search/distribution |
| Microsoft / Azure | Cloud + AI infrastructure | Enterprise AI lock-in | Azure, OpenAI partnership, government cloud |
| Amazon / AWS | Cloud infrastructure | Cloud dominance, Trainium chips | AWS, data center footprint, government cloud |
| Meta AI | Open-weight AI + social platforms | Ads, platform AI, open model influence | Llama ecosystem, social data, data centers |
| Oracle | Cloud and AI infrastructure | Hyperscale AI contracts | Cloud regions, enterprise/government clients |
| NVIDIA | AI chip leader | Sell accelerators, networking, software stack | GPUs, CUDA ecosystem, AI hardware roadmap |
| AMD / Intel | Chip challengers | Break NVIDIA dominance | CPUs, GPUs, accelerators |
| CoreWeave / Crusoe / Core Scientific | AI compute / specialized data center providers | Rent GPU capacity, build AI compute | GPU cloud capacity, energy-linked compute |
Data center operators such as Digital Realty, Equinix, QTS, CyrusOne, Vantage, Compass, CloudHQ, DataBank, Flexential, Iron Mountain, Aligned, and others control land, facilities, leasing, power contracts, cooling systems, uptime, and customer access. Their incentive is simple: secure cheap power, win long-term hyperscale leases, reduce permitting friction, and monetize the AI boom.
| Government player | Why it matters |
| White House | Executive orders, national-security framing, permitting priorities |
| Congress | Taxes, ownership proposals, moratoria, subsidies, privacy law |
| Department of Energy | Power, national labs, AI-science infrastructure, energy policy |
| Department of Defense / Army / Air Force | Military-base land, defense AI workloads, classified compute |
| Commerce / BIS | Export controls on chips and AI compute |
| CFIUS | Foreign ownership and national-security review |
| DHS / CISA | Critical infrastructure and cyber defense |
| NIST | AI risk frameworks and technical standards |
| FERC | Grid interconnection, transmission, wholesale power issues |
| State public utility commissions | Utility rates, grid upgrade cost allocation |
| Local zoning boards | Land use, permits, noise, water, local resistance |
5. Bernie Sanders / 50% Ownership Claim
Verdict: Verified as proposed legislation, not enacted law.
Sanders’ office says the American A.I. Sovereign Wealth Fund Act would give the public a direct ownership stake in the largest AI companies through a one-time 50% tax on stock. The stock would be placed into a sovereign wealth fund.
The bill summary states that the tax would apply to the largest AI companies and to new AI companies once they reach the threshold of $200 million in annual AI sales. It also says companies with both AI and non-AI businesses would have to separate those businesses so the public stake applies to the AI business.
The bill text is framed as an amendment to the Internal Revenue Code imposing an excise tax on systemically important AI activity, not as a classic eminent-domain seizure. But the tax is paid in equity, which makes it nationalization-adjacent in practical effect.
| Category | Classification |
| Enacted law | No |
| Introduced/proposed bill | Yes |
| 50% equity transfer | Yes, by tax paid in stock/equity |
| Government voting control | Yes, proposed through voting shares managed by commission |
| Direct nationalization | Not classic nationalization, but nationalization-adjacent |
| Applies to data centers | Potentially yes, if tied to AI data centers/AI computing infrastructure and statutory thresholds |
| Constitutional risk | High |
Constitutional pressure points include the Takings Clause, Due Process Clause, Equal Protection, bill-of-attainder concerns, First Amendment risks, and modern administrative-law limits. The government would likely frame the mechanism as a tax. Challengers would argue that a forced 50% equity transfer with voting consequences is functionally confiscatory and unlike ordinary revenue collection.
6. Trump / Data Centers as Military Installations Claim
Verdict: Not verified as a literal declaration. Verified as national-security framing and military-base leasing.
The strongest official evidence is Trump’s July 23, 2025 executive order, “Accelerating Federal Permitting of Data Center Infrastructure.” The order says AI data centers and infrastructure that powers them are essential to national security, economic prosperity, and scientific leadership; it prioritizes rapid buildout by easing federal regulatory burdens and using federally owned land/resources.
That is not the same thing as declaring every data center a military installation. However, the military-adjacent angle is real. The Army announced conditional agreements involving hyperscale data centers at Fort Bliss and Dugway Proving Ground, with the Army Corps of Engineers involved in lease negotiations and environmental review. The Air Force also opened unused land on five bases for private AI data centers.
| Claim | Verdict |
| Trump declared data centers military bases | Not verified / misleading |
| Trump treated AI data centers as national-security infrastructure | Verified |
| Trump accelerated federal permitting for AI data centers | Verified |
| Federal land may be used for AI data centers | Verified |
| Military bases may host private/commercial AI data centers | Verified |
| All private data centers are now military installations | False based on available evidence |
| Data centers are military-adjacent assets | Strongly supported |
A private data center can be critical infrastructure, a defense contractor facility, a facility on federal or military land, a protected site under contract, part of the defense industrial base, or a privately owned facility with classified workloads. Those are not all the same thing as a military installation.
7. Energy, Water, and Grid Pressure
Data centers are now a grid issue. The Lawrence Berkeley National Laboratory 2024 U.S. Data Center Energy Usage Report says U.S. data center electricity use rose from 58 TWh in 2014 to 176 TWh in 2023 and could reach 325-580 TWh by 2028. It also estimates data centers consumed about 4.4% of U.S. electricity in 2023 and could reach 6.7% to 12% by 2028, depending on growth.
The International Energy Agency projects global data center electricity consumption will roughly double by 2030, with AI-focused data centers growing faster than overall data center demand. The U.S. Department of Energy cites EPRI’s estimate that data centers could consume up to 9% of U.S. electricity generation annually by 2030.
| Issue | What matters |
| Electricity load | AI facilities can demand hundreds of MW or more. |
| Transmission bottlenecks | Power may exist regionally but cannot physically reach the site. |
| Ratepayer fairness | If utilities socialize grid upgrades, households and small businesses may pay part of the cost. |
| Nuclear deals | AI firms increasingly seek firm clean power, including nuclear. |
| Natural gas buildout | Fast reliable power often means gas turbines, which creates emissions fights. |
| Backup diesel | Data centers use diesel generators for reliability, raising local emissions concerns. |
| Water | Evaporative cooling can reduce electricity use but may stress local water systems. |
| Noise | Cooling fans, backup generators, and substations can trigger local opposition. |
| Jobs | Construction jobs may be large; permanent operations jobs are usually much smaller than political sales pitches imply. |
This is where citizens feel the AI boom first - not in philosophical debates about AGI, but in utility bills, zoning meetings, water fights, and substations.
8. Legal and Constitutional Issues
Federal ownership stakes: A forced 50% equity transfer framed as a tax would trigger hard constitutional fights. Congress has broad taxing power, but a tax that requires stock transfer and voting-control consequences is not a normal revenue measure. The closer it looks to confiscation or forced public ownership, the more vulnerable it becomes.
| Case / doctrine | Why it matters |
| Horne v. Department of Agriculture | Government taking of personal property can require just compensation; not only land gets Takings Clause protection. |
| Cedar Point Nursery v. Hassid | Physical access mandates can be takings. |
| Tyler v. Hennepin County | Government cannot keep more property value than it is owed without takings concerns. |
| Sheetz v. County of El Dorado | Legislative permit exactions are not immune from Takings Clause scrutiny. |
| Penn Central | Regulatory takings balancing test. |
| Lucas | Total economic wipeout can be a taking. |
| Kelo | Broad public-use doctrine, but politically controversial. |
| Loper Bright | Courts no longer defer to agency statutory interpretations simply because a statute is ambiguous. |
| Jarkesy | Limits agency use of in-house tribunals when civil penalties trigger Seventh Amendment rights. |
| Youngstown Sheet & Tube | Even emergency claims do not give the President unlimited power to seize or control private industry without congressional authority. |
| NRA v. Vullo | Government officials cannot use coercive regulatory threats to suppress speech through private intermediaries. |
Privacy and surveillance concerns depend on legal process and context. Cloud infrastructure stores communications, files, metadata, business records, location data, health data, financial data, and government data. Government access usually requires subpoenas, warrants, FISA orders, national-security letters, or statutory process depending on the data and context. Carpenter, Riley, Katz, Jones, and Kyllo all matter because they show the Supreme Court struggling to translate the Fourth Amendment into digital infrastructure.
9. National Security and Military Angle
AI data centers can realistically be strategic assets. That part is not conspiracy theory. The national-security case is strong in military logistics, cyber defense, intelligence analysis, satellite imagery, drone coordination, battlefield decision support, autonomous systems, classified cloud workloads, defense contractor AI, code analysis, biosecurity modeling, weapons simulation, and supply-chain monitoring.
The danger is mission creep. “National security” can mean legitimate protection from China, cyberattacks, sabotage, and supply-chain dependency. It can also become a political skeleton key that unlocks fast permits, secrecy, local bypass, public subsidy, private profit, and reduced accountability.
10. Surveillance, Censorship, and Civil Liberties
| Concern | Classification | Why |
| Government access to cloud data through legal process | Documented | Warrants, subpoenas, FISA, and national-security letters exist as legal pathways. |
| AI moderation systems shaping speech visibility | Documented / plausible | Platforms and AI providers already use moderation systems; government pressure raises constitutional issues. |
| Public-private censorship pressure | Documented legal risk | NRA v. Vullo confirms coercive government pressure on private actors can violate the First Amendment. |
| AI-driven propaganda and synthetic media | Technically possible / documented in broader AI use | The tools exist; attribution and intent require case-specific evidence. |
| Digital identity linked to AI access | Plausible but not universal | Some policy proposals move this way; not broadly enacted as a national AI access requirement. |
| Predictive policing / biometric surveillance | Documented in pieces | Exists in law enforcement technology markets; expansion depends on jurisdiction and procurement. |
| Government decides what reality AI can show | Plausible risk, not proven as a unified program | Would require evidence of formal control, coercive pressure, procurement terms, or regulatory mandates. |
The cleanest constitutional line: government may speak, advise, regulate within lawful bounds, and enforce valid law. Government may not use threats, licensing pressure, funding leverage, or regulatory coercion to turn private AI/cloud companies into speech-control deputies.
11. Follow the Money
Compute demand: AI labs and enterprises need GPUs and cloud capacity.
Real estate and power: Data center developers monetize land, power contracts, and long leases.
Finance: Private equity, REITs, sovereign wealth funds, and infrastructure funds want stable long-term returns.
Government leverage: Subsidies, tax abatements, permitting, federal land, defense contracts, and energy policy shape who wins.
The Army’s data center initiative shows the new model clearly: private companies build and operate commercial hyperscale data centers on military land, while the Army gains access to compute capacity and infrastructure benefits.
The honest question is not “Are data centers good or bad?” The honest question is: who pays for the power, who gets the tax break, who owns the asset, and who controls the switch?
12. Local Community Impact
Zoning changes and land acquisition
Noise from cooling systems, backup generators, and substations
Water use and cooling demands
Diesel backup emissions
Property-value concerns
Secrecy around site selection
Tax incentive deals
Emergency services burden
Road and construction impacts
Electric-rate increases
Limited permanent jobs compared with construction jobs
Reduced local control if state or federal officials fast-track projects
This is why data center fights are becoming bipartisan at the local level. One side may object because of climate and water. Another may object because of property rights, secrecy, Chinese competition, surveillance, or local sovereignty. Different roads, same zoning meeting.
13. Data Centers as the New Military-Industrial-Tech Complex
The old military-industrial complex was defense contractors, weapons systems, bases, Congress, and procurement. The new version adds Big Tech, cloud providers, AI labs, chip companies, data center REITs, private equity, utilities, nuclear and gas developers, intelligence agencies, universities, state economic-development offices, and federal AI regulators.
Power now flows like this: citizens generate data -> AI companies train models -> cloud/data centers run systems -> utilities power compute -> government buys, uses, or regulates AI -> finance capitalizes expansion -> local communities absorb land, water, noise, and grid impacts.
That does not mean every piece is corrupt. It means the structure is powerful enough that normal citizens can lose visibility fast.
14. Claim-by-Claim Fact Check
| Claim | Verdict | Evidence |
| Bernie Sanders wants the government/public to own 50% of major AI companies. | Verified as proposal | Sanders’ bill summary and press release describe a one-time 50% stock/equity tax into a sovereign wealth fund. |
| This is enacted law. | False | Evidence shows proposal/bill, not enacted law. |
| The proposal includes voting shares. | Verified | Sanders’ press release says the commission would use voting shares to block harmful decisions and push preferred policies. |
| The proposal applies only to OpenAI. | False | It covers qualifying AI companies and includes AI data centers, AI computing infrastructure, AI services, and advanced robotics. |
| Trump declared data centers military bases. | Misleading / unsupported | EO uses national-security framing and federal-land policy; military branches are leasing base land, but no blanket legal declaration was found. |
| AI data centers are becoming national-security assets. | Strongly supported | Trump EO explicitly ties AI data centers and power infrastructure to national security; military branches are pursuing AI data center projects on bases. |
| The government wants to own data centers. | Plausible but not proven as broad policy | Sanders proposal may reach AI data center/compute companies; Trump policy focuses on permitting/federal land, not ownership. |
| Data centers can override local zoning. | Legally possible in some contexts, not blanket enacted rule | Federal land and federal permitting can reduce local leverage; state/local authority still matters off federal land. |
| AI data centers will crash the grid. | Speculative as blanket claim | Load growth is real and regionally serious; collapse is not established. |
| Citizens may pay higher power bills because of AI. | Plausible / region-dependent | Depends on utility rate design, upgrade cost allocation, generation constraints, and state PUC decisions. |
| AI infrastructure is the new oil. | Metaphorically strong, literally imperfect | Compute is becoming strategic like oil was, but it is not a commodity fuel; it is a stack of chips, power, software, data, and access. |
| Big Tech and government are merging. | Overstated but directionally supported | Public-private partnerships, defense cloud contracts, federal land leasing, and regulation create tighter integration. |
| This is all conspiracy theory. | False | The ownership proposal, federal permitting EO, and military-base data center projects are documented. |
| This is nationalization. | Partly misleading | Sanders’ bill is not classic seizure, but a forced 50% equity tax with voting shares is nationalization-adjacent. |
15. Red Flags and Green Flags
Red flags
Forced government equity stakes in private AI firms
Government voting shares or board-like control
National-security language with no limiting principle
Emergency powers with no sunset clause
Bypassing local zoning without local process
Secret utility deals and hidden ratepayer subsidies
Public-private censorship pressure
Unrestricted government access to cloud or AI data
AI model licensing controlled by agencies with vague standards
Civilian infrastructure treated as military-adjacent without transparency
Federal land leases that privatize profit while socializing power/water risk
No judicial review or weak due process for affected communities
Green flags
Clear statutory authority
Public hearings
Local consent
Transparent subsidy agreements
Ratepayer protection
Water-use disclosure
Environmental review
No government voting control over speech-sensitive AI systems
Warrant requirements for data access
Sunset clauses
Independent audits
Clear wall between civilian AI infrastructure and military/classified use
16. Source Ranking
Highest-value sources for this issue:
Bill text and bill summaries
Executive orders and White House fact sheets
Congressional Record and committee materials
Federal Register notices
Agency solicitations and lease documents
Utility commission filings
FERC records
SEC filings and investor disclosures
Local zoning and tax-abatement agreements
Government contract databases
CRS, GAO, DOE, LBNL, IEA, EPRI, NIST
Court filings and judicial opinions
Serious financial, energy, and technology reporting
Use cautiously: campaign speeches, viral clips, partisan commentary, company press releases, advocacy-group claims, articles that cite other articles in circles, and social-media posts claiming “military bases” or “nationalization” without documents.
17. Final Bottom-Line Conclusion
What is actually happening is this: AI compute has crossed from tech industry issue into national power issue. The fight is no longer only about apps, chatbots, or Silicon Valley billionaires. It is about chips, electricity, water, land, military use, censorship risk, ownership, and constitutional limits.
Confirmed: Sanders proposed a 50% equity-stock tax on large/systemically important AI companies through a sovereign wealth fund. Trump signed an executive order accelerating AI data center permitting and tying AI infrastructure to national security. The Army and Air Force are pursuing private/commercial AI data center projects on military bases or military land.
Proposed: Public ownership stakes, AI data center moratoria, accelerated federal permitting, federal land leasing, and military-adjacent AI compute partnerships.
Exaggerated: “Trump declared all data centers military installations.” The evidence supports national-security framing and military-base leasing - not a blanket legal conversion of private data centers into military bases.
Legally possible: Critical-infrastructure designation, defense contracts, federal land leases, export controls, tax incentives, emergency-power actions, utility regulation, and possibly carefully designed public equity structures.
Constitutionally dangerous: Forced equity transfer with voting rights, vague emergency control, public-private speech coercion, broad agency licensing of AI models, warrantless access to cloud data, and federal preemption that wipes out local property rights without due process.
Who benefits: Big Tech, AI labs, chipmakers, data center developers, utilities, private equity, defense planners, and politicians who gain leverage over the future economy.
Who pays: Ratepayers, local communities, water systems, taxpayers through subsidies, and citizens if civil-liberties lines are blurred.
The cleanest one-sentence read: The fight is not just who owns the servers. It is who owns the compute, who controls the power, who writes the rules, who gets the data, and who decides what reality the machine is allowed to show us.
Appendix A. Key Source Notes
Bernie Sanders, “Sanders Introduces Legislation to Create $7 Trillion AI Sovereign Wealth Fund,” official Senate press release, https://www.sanders.senate.gov/press-releases/news-sanders-introduces-legislation-to-create-7-trillion-ai-sovereign-wealth-fund/
American A.I. Sovereign Wealth Fund Act Summary, official Sanders Senate PDF, https://www.sanders.senate.gov/wp-content/uploads/AmericanAISovereignWealthFundActSummary.pdf
American A.I. Wealth Fund bill text, official Sanders Senate PDF, https://www.sanders.senate.gov/wp-content/uploads/AmericanAIWealthFundTextv618.pdf
White House, Executive Order: “Accelerating Federal Permitting of Data Center Infrastructure,” July 23, 2025, https://www.whitehouse.gov/presidential-actions/2025/07/accelerating-federal-permitting-of-data-center-infrastructure/
U.S. Army, “Army reaches conditional agreement with private industry for hyperscaled data centers,” https://www.army.mil/article/291360/army_reaches_conditional_agreement_with_private_industry_for_hyperscaled_data_centers
Federal News Network, Air Force unused land on military bases for private AI centers, https://federalnewsnetwork.com/federal-newscast/2025/10/air-force-opening-up-unused-land-on-five-military-bases-for-private-ai-centers/
Lawrence Berkeley National Laboratory, U.S. Data Center Energy Usage Report coverage, https://bies.lbl.gov/news/berkeley-lab-report-evaluates-increase-electricity-demand-data-centers
International Energy Agency, Key questions on energy and AI, https://www.iea.org/reports/key-questions-on-energy-and-ai/executive-summary
U.S. Department of Energy, clean energy resources to meet data center electricity demand, https://www.energy.gov/oe/clean-energy-resources-meet-data-center-electricity-demand
Supreme Court of the United States, Loper Bright Enterprises v. Raimondo, https://www.supremecourt.gov/opinions/23pdf/22-451_7m58.pdf
Supreme Court of the United States, SEC v. Jarkesy, https://www.supremecourt.gov/opinions/23pdf/22-859_1924.pdf
Supreme Court of the United States, NRA v. Vullo, https://www.supremecourt.gov/opinions/23pdf/22-842_6kg7.pdf
Supreme Court / Justia, Sheetz v. County of El Dorado, https://supreme.justia.com/cases/federal/us/601/22-1074/
White House, “Promoting Advanced Artificial Intelligence Innovation and Security,” June 2026, https://www.whitehouse.gov/presidential-actions/2026/06/promoting-advanced-artificial-intelligence-innovation-and-security/