WHO'S ON FIRST: THE PRIVATIZATION OF AMERICAN DEMOCRACY
When Billionaires Inherit the Earth — and the Satellites, the Classrooms, and the Chain of Command
"Who's on first?" "Yes." "I mean the fellow's name." "Who." "The guy on first." "Who." "The first baseman." "Who is on first." — Abbott and Costello, 1938
"Who controls the satellite?" "Elon." "Who authorized that?" "Unclear." "Who's accountable?" "...Who." — American democracy, 2026
There is a peculiar kind of comedy unfolding in the theater of American governance — the kind that isn't funny at all once you read the fine print. A small cohort of extraordinarily wealthy men have, through the perfectly legal accumulation of essential infrastructure, acquired something that no election conferred and no constitution anticipated: the practical power to govern without the inconvenience of being elected. Call it the unelected veto. Call it technocratic feudalism. Call it what it is — a structural transformation of democratic sovereignty so gradual, so wrapped in the language of innovation and efficiency, that we nearly missed it while updating our apps.
This is not a conspiracy theory. It is a procurement problem with geopolitical consequences, an education crisis dressed in Silicon Valley branding, and a democratic accountability gap wide enough to fly a restricted Ukrainian drone through — assuming, of course, the satellite is cooperating that day.
Part I: Who Voted for Elon's Foreign Policy?
Here is a fact so strange it deserves its own dramatic pause: In October 2026, the trajectory of a war in Eastern Europe was briefly altered by velocity-based restrictions placed on commercial satellite terminals — restrictions that limited drone connectivity above roughly 100 mph, disrupting Ukrainian long-range reconnaissance and strike operations along active front lines.
Ukrainian commanders described losing drones mid-mission. Communications cut out. The battlefield went partially dark.
Nobody voted for this. No Senate confirmed it. No treaty authorized it.
SpaceX's Starlink — the commercial service — had previously been restricted by Musk citing terms-of-service violations when Ukraine used civilian terminals for offensive drone operations. Musk's stated position on Starshield, the military-grade variant, was more nuanced: he argued those systems are owned and controlled by the U.S. government, with usage parameters set by the Pentagon and the President, not SpaceX.
The October restrictions, however, landed in disputed territory — literally and figuratively. Ukrainian officials reported the impact in stark operational terms. Diplomatic pressure followed. After a week of frantic back-channel negotiations between Kyiv, Washington, and SpaceX, full terminal capabilities — connectivity at speeds exceeding 400 km/h — were restored.
Crisis averted. Lesson unlearned.
Because here is the structural problem that no firmware update can patch: a private executive's operational decision, commercial risk tolerance, or personal foreign-policy instinct can now constrain what a sovereign military is able to do in real time. That is not a bug in the Starlink terms of service. That is a fundamental reordering of who holds power.
The Unelected Veto, Defined
The phrase does not mean a tech billionaire has acquired a formal constitutional veto — the kind that requires a two-thirds override. It means something more insidious: control over essential technical infrastructure that the government cannot immediately replace gives a private actor leverage over decisions with public consequences.
Consider the contrast:
| Traditional Public Authority | Private Infrastructure Dependency |
|---|---|
| Government sets military objectives | Commercial service restrictions can constrain how those objectives are pursued |
| Commanders direct operations through lawful chain of command | Commanders may depend on a provider's network availability and technical rules |
| Elected officials are accountable for public decisions | A private executive is accountable to shareholders, contracts, and corporate governance |
| Government can order its own personnel to act | It may lack the immediate ability to compel a private network to deliver a capability |
The important point — and it bears repeating with the emphasis of a constitutional law professor who has just spilled their coffee — is that this is a governance problem regardless of which executive, company, or administration is involved. It is not about Elon Musk's politics. It is about what happens when the architecture of power allows a private company to sit between a government and its own lawful decisions.
Part II: The Dependency Trap — How Governments Hollow Themselves Out
The Starlink episode is dramatic because it involves missiles and drones and the fog of war. But the same structural problem exists in considerably less cinematic settings: federal cloud contracts, software licensing agreements, and the IT systems that run the daily operations of American governance.
The U.S. Government Accountability Office has been quietly documenting this for years, in the patient, methodical tone of an accountant who has discovered the building is on fire and is carefully noting it in a spreadsheet.
- 17 of 24 federal agencies reported cloud-cost management challenges in a June 2026 GAO report.
- 15 agencies reported procurement difficulties caused by outdated regulations.
- 11 agencies struggled to implement multi-vendor cloud solutions — meaning they are, in practice, locked in.
- Federal agencies spend more than $100 billion annually on IT, with roughly 80% historically going toward maintaining existing legacy systems rather than building new capacity.
- 5 of 6 agencies surveyed encountered restrictive software licensing practices — including fees for using software with competing cloud providers.
This is the slow version of the unelected veto. No dramatic satellite cutoff. Just the gradual, compounding reality that a government which cannot afford to switch providers, cannot rebuild its own technical capacity, and cannot audit the systems it depends upon has quietly transferred operational authority to whoever holds the contract.
The Four Stages of Dependency
The progression follows a predictable arc:
- Outsourcing — A private vendor supplies a specialized service faster or cheaper than the government can build it.
- Integration — Public operations, data, workflows, and staff become deeply embedded in the vendor's systems.
- Dependency — The government loses the people, expertise, infrastructure, or budget needed to switch or operate independently.
- Reduced public control — The vendor's pricing, policies, service restrictions, or technical decisions can now materially constrain public action.
The dependency becomes self-reinforcing: the more a government relies on one system, the more expensive it becomes to leave, and the harder it becomes to develop a replacement. This is not a conspiracy. It is a procurement cycle. It is also, at sufficient scale, a sovereignty problem.
Part III: Corporate Capture of the Classroom — Who Controls What Children Learn?
If the satellite story is about war, and the procurement story is about governance, the education story is about something more foundational still: the formation of the next generation of democratic citizens.
Artificial intelligence is transforming public education with genuine promise. AI tutors can personalize instruction, expand access, reduce teacher administrative burden, and reach students who need additional support. These are real benefits. They deserve acknowledgment before the critique, because the critique is serious.
Here is the problem: when public schools rely on privately owned AI systems to teach, assess, and guide students, who controls the educational experience — and who is accountable for its consequences?
The Shift from Public Good to Proprietary Product
Public education, in democratic theory, exists to develop informed citizens capable of critical thought and civic participation. It is a public good — universal, accountable, and oriented toward the common interest.
Commercial AI platforms operate under a different set of incentives:
| Commercial Mechanism | Potential Educational Consequence |
|---|---|
| Subscription-based platforms | Schools become dependent on recurring payments to maintain access |
| Proprietary instructional systems | Teachers find it difficult to move lessons, records, and workflows to another provider |
| Automated assessment | Students evaluated using criteria that teachers and families cannot fully inspect |
| Student analytics | Schools collect more behavioral and performance data than necessary |
| AI-generated instructional content | Vendor defaults influence which explanations, examples, and perspectives students encounter |
| Bundled software ecosystems | Purchasing one product creates dependence on the same company's broader infrastructure |
The crucial distinction is between purchasing a tool and allowing the tool to become the operating system of public education. A school district that buys software while retaining control over curriculum, data, assessment, and teacher judgment is in a fundamentally different position from one that has outsourced all four.
Algorithmic Bias and Epistemic Control
AI systems learn from data. Data is not neutral. It reflects the world that produced it — its historical inequities, its dominant linguistic norms, its cultural assumptions, its gaps and silences.
When an AI writing assessment consistently interprets one dialect as less sophisticated than another, that is not a technical glitch. It is a value judgment encoded in software, administered with the authority of an institution. When an AI history tutor provides rich, detailed answers about some communities and superficial responses about others, that asymmetry shapes what students understand to be historically significant.
The democratic danger is not merely that an algorithm may be wrong. It is that its judgment may acquire the authority of a public institution without the scrutiny normally expected of that institution.
UNESCO's guidance on generative AI in education specifically identifies concerns about privacy, cultural and linguistic diversity, the homogenization of opinions, equity, and human agency. A February 2025 RAND report found that roughly one in four U.S. K–12 teachers used AI tools for instructional planning — while cautioning that adoption statistics do not establish educational improvement. A March 2026 RAND report found growing student concerns about AI's effects on critical thinking.
California issued statewide AI-in-education guidance in January 2026, addressing data privacy, equity, academic integrity, and human-centered learning. It is a useful framework. The harder question is whether districts have the staffing, funding, and procurement expertise to implement it.
The Fundamental Governance Question
If artificial intelligence becomes part of the educational infrastructure of a democratic society, who should decide how it works, what values it reflects, what data it uses, and what authority it possesses?
This is not a rhetorical question. It is the most important question in education policy right now, and it is being answered — by default, through purchasing decisions — in favor of whoever holds the contract.
Part IV: Democracy in the Age of Technocratic Feudalism
The term technocratic feudalism is a metaphor, not a legal designation. But metaphors earn their keep when they illuminate something true about power.
| Medieval Feudalism | Technocratic Feudalism (The Concern) |
|---|---|
| Lords controlled land — the essential productive resource | Platforms and tech firms control digital infrastructure — the essential communicative resource |
| Subjects depended on the lord's estate to survive | Users, businesses, and governments depend on digital ecosystems to function |
| Power stemmed from ownership of land and military capacity | Power stems from ownership of data, networks, computation, and connectivity |
| Authority was geographically concentrated | Influence is concentrated among a small number of global firms |
| Accountability ran upward to the lord, not outward to subjects | Accountability runs to shareholders and boards, not to the public |
The analogy is imperfect — modern citizens possess legal rights that medieval serfs did not, and democratic governments retain the formal authority to regulate, break up, and legislate. But the metaphor is useful precisely because it directs attention toward the right questions: Who owns essential resources? Who sets the rules of access? Who can challenge those rules? And who has the power to change them?
The Information Monopoly
Democracy requires more than the right to vote. It requires access to information, the ability to criticize public officials, and a reasonable chance of having one's concerns heard.
A Pew Research Center survey from August 2026 found that 40% of U.S. adults regularly get news from Facebook and 34% from YouTube. These platforms do not own newspapers. They do not need to. Their recommendation algorithms, search rankings, moderation policies, and advertising systems can determine which information reaches which people — without writing a single headline.
Public Citizen's analysis of Federal Election Commission disclosures reported $646 million in disclosed federal corporate political spending in the 2026 midterm cycle, with crypto, online betting, Big Tech, AI, and data-center-related sectors accounting for 53% of that total. Issue One reported that 11 major technology and AI companies and their leading trade associations spent a combined $41 million lobbying the federal government in the first half of 2026 alone.
These figures do not prove that elections were purchased or that every expenditure produced political compliance. They establish something more specific: the industries most dependent on favorable AI, data, and platform regulation have deployed resources to influence that regulation at a scale that ordinary citizens cannot match.
When access and influence depend heavily on wealth, formal political equality can coexist with profound inequalities in practical political power.
Part V: What Democratic Resilience Actually Looks Like
The answer to technocratic feudalism is not Luddism, nationalization of every server farm, or the romantic fantasy of a government that builds its own satellites. It is something more precise and more demanding: ensuring that outsourcing does not eliminate the public sector's ability to act.
For Government Infrastructure
| Safeguard | What It Accomplishes |
|---|---|
| Government retains critical technical expertise | Officials can independently evaluate providers and make informed decisions |
| Multiple providers and interoperable systems | Reduces reliance on a single company where technically feasible |
| Data portability and documented exit plans | Makes switching providers or recovering operations practical |
| Clear service and emergency-control provisions | Establishes who can restrict services, under what conditions, and with what notice |
| Independent audits and legislative oversight | Makes contracts, risks, and performance accountable to the public |
| Government-controlled alternatives for essential services | Preserves continuity when commercial systems fail or become unavailable |
| Regular contingency exercises | Tests whether the government can actually operate if a provider disappears |
For Education
Public schools should demand: public control of educational decisions (teachers and qualified officials retain responsibility for curriculum, grading, and placement); independent evaluation of AI tools with testing for disparate impacts; strong student-data protections with data minimization and enforceable contractual safeguards; interoperability and exit rights so districts can change vendors without catastrophic disruption; and investment in public-sector technical capacity so schools can assess products rather than simply accept vendor claims.
For Democracy Broadly
- Campaign-finance transparency — Citizens should be able to identify major financial interests seeking to influence public policy.
- Competition and interoperability — Preserve viable alternatives to dominant platforms and make switching practical.
- Independent oversight of critical infrastructure — Require contingency plans, auditable contracts, and clear rules for essential government services.
- A pluralistic information environment — Protect independent journalism, support media literacy, and preserve space for competing sources of information.
- Meaningful public participation — Ensure communities can scrutinize major technology projects, question public subsidies, and challenge consequential policies through democratic processes.
Conclusion: Abbott, Costello, and the Architecture of Power
"Who's on first?"
In 1938, the joke was about the absurdity of miscommunication. In 2026, the joke — if we can still call it that — is about the absurdity of a system in which the answer to nearly every question about who controls essential public infrastructure is: a private company you didn't elect, can't fire, and may not be able to replace.
Who controls the satellite network that Ukraine's military depends on? A private company. Who controls the cloud infrastructure that 17 federal agencies can't afford to leave? A handful of private companies. Who controls the AI systems shaping how the next generation of Americans learns to think? Private companies. Who controls the platforms through which 40% of Americans get their news? Private companies. Who authorized all of this? Nobody, exactly. It accumulated. It was procured. It was efficient.
The privatization of sovereign public functions is not a conspiracy. It is the logical endpoint of decades of outsourcing without adequate attention to what was being surrendered along with the contract. Governments retained the legal authority. They transferred the practical capacity. The result is a system in which elected officials can be held responsible for outcomes they no longer fully control, while the private actors who do control them answer primarily to shareholders and terms of service.
That is the real unelected veto: not a formal power, but a structural condition — the gap between who is accountable and who is in control.
The democratic test, in the end, is straightforward: Can citizens obtain information from multiple sources, organize independently, influence policy, and hold decision-makers accountable? Can governments carry out their lawful responsibilities without depending on the discretionary cooperation of a private provider? Can communities meaningfully shape the technology that affects their lives?
A democracy remains resilient when access to essential systems does not depend on the favor of a powerful individual — when public institutions retain the capacity to act, and when citizens have meaningful ways to challenge decisions that affect them.
The infrastructure of democracy should not be a subscription service. It should not have a terms-of-service agreement. And it should not cut out at 100 miles per hour.
The first baseman is still "Who." The difference is that in 2026, "Who" owns the stadium, the broadcast rights, the ticketing platform, and the satellite uplink. And nobody voted for that.
Sources & Further Reading
Who's on First: The Privatization of American Democracy
🛰️ I. Starlink, Starshield & Ukraine — The Unelected Veto in Action
The October 2026 satellite restriction crisis generated some of the most important primary reporting on private infrastructure and sovereign military operations.
Reuters — "Order to limit Starshield for Ukraine came from inside White House" (October 15, 2026) Speed restrictions capped at 100 mph; sourcing traces the order to the White House rather than SpaceX unilaterally. 🔗 https://www.reuters.com/world/europe/order-limit-starshield-ukraine-came-inside-white-house-sources-say-2026-10-15/
Reuters — "What is Starshield? SpaceX's military version of Starlink" (October 23, 2026) Explains the technical and contractual distinction between commercial Starlink and military Starshield terminals, and the operational impact on Ukrainian defense. 🔗 https://www.reuters.com/business/media-telecom/what-is-starshield-spacexs-military-version-starlink-2026-10-23/
Reuters — "Ukraine says Starlink speed curbs remain despite Trump assurance" (October 23, 2026) Documents the contradiction between U.S. government statements and Ukrainian field reports during the restriction period. 🔗 https://www.reuters.com/business/media-telecom/ukraine-says-starlink-speed-curbs-remain-despite-trump-assurance-2026-10-23/
Reuters — "Ukraine says Starshield back to normal, Musk says Kyiv operating at 400 kph" (October 24, 2026) Confirms restoration of full terminal capabilities following diplomatic pressure; documents the week-long disruption. 🔗 https://www.reuters.com/world/europe/ukraine-says-starlink-starshield-operating-400-kph-after-week-throttling-2026-10-24/
🏛️ II. Government IT Dependency — The GAO Reports
The Government Accountability Office has produced the most authoritative documentation of federal dependency on private technology vendors.
U.S. GAO — "Cloud Computing: Federal Government Needs to Address Procurement Challenges" (GAO-26-107530, June 2026) The primary source for the finding that 17 of 24 agencies reported cloud-cost challenges, 15 reported procurement difficulties, and 11 struggled with multi-vendor implementation. 🔗 https://www.gao.gov/products/gao-26-107530
U.S. GAO — "Cloud Computing: Selected Agencies Need to Implement Vendor Lock-In Strategies" (GAO-25-107114) Documents specific vendor lock-in practices, including hardware restrictions preventing competing software use. 🔗 https://www.gao.gov/products/gao-25-107114
U.S. GAO — "Agencies Need to Plan for Modernizing Critical Decades-Old IT Systems" (GAO-25-107795, 2025) Source for the finding that federal agencies spend ~$100 billion annually on IT, with ~80% going to maintaining legacy systems; only 3 of 11 critical modernizations completed. 🔗 https://www.gao.gov/products/gao-25-107795
FedScoop — "The GAO flagged 10 'critical' legacy IT systems. Years later, most haven't been modernized" Accessible journalism summarizing the GAO legacy IT findings, including staff reductions and contract cuts affecting modernization timelines. 🔗 https://fedscoop.com/the-gao-flagged-10-critical-legacy-it-systems-years-later-most-havent-been-modernized/
💰 III. Corporate Political Spending & Democratic Capture
The 2026 midterm cycle produced landmark data on technology-sector political spending.
Public Citizen — "New Report Exposes Unprecedented Corporate Spending in 2026 Midterm Elections" Primary source for the $646 million figure in disclosed federal corporate political spending; documents the 40% increase over 2024. 🔗 https://www.citizen.org/news/new-report-exposes-unprecedented-corporate-spending-in-2026-midterm-elections/
Public Citizen — "Corporate Money in the 2026 Midterms Surges to $650 Million" Full analytical report with sector-by-sector breakdown; crypto, AI, Big Tech, and online betting sectors account for 53–57% of total corporate midterm spending. 🔗 https://www.citizen.org/article/corporate-money-2026-midterms-surges-to-650-million/
Public Citizen — "The Rise of Corporate Supremacist Super PACs" (Full PDF Report) Detailed breakdown of $294 million in combined crypto, AI, Big Tech, and online betting corporate spending; documents the emergence of single-industry Super PACs with outsized electoral influence. 🔗 https://www.citizen.org/wp-content/uploads/corporate-supremacist-super-pacs-crypto-ai-2026-midterms.pdf
📚 IV. AI in Education — Corporate Capture of the Classroom
UNESCO — "Guidance for Generative AI in Education and Research" The foundational international policy document on AI in education; addresses privacy, cultural diversity, homogenization of opinion, equity, and the need for human-centered safeguards. 🔗 https://www.unesco.org/en/articles/guidance-generative-ai-education-and-research
RAND Corporation — "Teachers' Use of AI in K–12 Education" (February 2025) Source for the finding that ~1 in 4 U.S. K–12 teachers used AI tools in 2023–24; cautions that adoption statistics do not establish educational improvement. 🔗 https://www.rand.org/pubs/research_reports/RRA956-21.html
California Department of Education — "Artificial Intelligence Guidance for California Public Schools" (January 9, 2026) The statewide framework addressing student data privacy, equity, academic integrity, and human-centered learning in AI deployment. 🔗 https://www.cde.ca.gov/ls/ai/
White House Executive Order on AI Education (April 2025) Directed federal agencies to pursue public-private partnerships and identify funding mechanisms for AI instruction in American schools; the policy backdrop for accelerating EdTech adoption. 🔗 https://www.whitehouse.gov/presidential-actions/2025/04/advancing-artificial-intelligence-education-for-american-youth/
🌐 V. Technocratic Feudalism, Platform Power & Democratic Theory
Brookings Institution — "Tech Industry Political Advocacy and Data Center Infrastructure" (July 2026) Analyzes how technology-industry leaders are investing heavily in political advocacy as local communities challenge data-center projects; documents the intersection of private infrastructure investment and democratic decision-making. 🔗 https://www.brookings.edu
Pew Research Center — "News Platform Fact Sheet" (August 2026) Source for the finding that 40% of U.S. adults regularly get news from Facebook and 34% from YouTube; essential data on platform dependency in the public information environment. 🔗 https://www.pewresearch.org/journalism/fact-sheet/news-platform-fact-sheet/
Issue One — "Big Tech Lobbying Report, First Half 2026" Documents $41 million in combined lobbying expenditures by 11 major technology, social media, and AI companies and their trade associations in the first half of 2026. 🔗 https://www.issueone.org
Carnegie Endowment for International Peace — "Private Tech Companies and Sovereign Military Operations" Scholarly analysis of how private technology firms have acquired quasi-sovereign influence over state military and foreign-policy operations; essential theoretical grounding for the "unelected veto" concept. 🔗 https://carnegieendowment.org
🔬 VI. Academic & Long-Form Further Reading
For readers who want to go deeper into the theoretical and policy dimensions:
Shoshana Zuboff — The Age of Surveillance Capitalism (2019) The foundational text on how behavioral data extracted from users becomes the raw material for a new economic logic that threatens democratic autonomy.
Mariana Mazzucato — The Entrepreneurial State (2013, updated 2018) Argues that public investment underlies most major technological innovation, and that the privatization of those returns represents a structural transfer of public wealth to private actors.
Frank Pasquale — The Black Box Society (2015) Examines how secret algorithms in finance, search, and reputation systems concentrate power and undermine accountability — directly relevant to AI in education and platform governance.
Lina Khan — "Amazon's Antitrust Paradox", Yale Law Journal (2017) The paper that redefined antitrust thinking for the platform era; argues that consumer-welfare frameworks fail to capture the structural power of digital monopolies. 🔗 https://www.yalelawjournal.org/note/amazons-antitrust-paradox
🗂️ Quick Reference Summary
| Theme | Key Source | Citation |
|---|---|---|
| Starshield restrictions origin | Reuters, Oct. 15, 2026 | |
| Service restored at 400 kph | Reuters, Oct. 24, 2026 | |
| Federal cloud vendor lock-in | GAO-26-107530 | |
| Legacy IT modernization failures | GAO-25-107795 | |
| $646M corporate midterm spending | Public Citizen, 2026 | |
| AI in education guidance | UNESCO | |
| Teacher AI adoption data | RAND, Feb. 2025 | |
| California AI education policy | CA Dept. of Education | |
| News platform dependency | Pew Research, Aug. 2026 | |
| Platform power theory | Zuboff, Surveillance Capitalism |
All URLs verified as of October 25, 2026. Institutional homepages are listed where specific report URLs require database access or login. Academic texts are available through university libraries and major booksellers.
