DELIVER AMERICA

A Proposal for the Public Acquisition of Amazon’s Delivery Network and the Transformation of the United States Postal Service

EXECUTIVE SUMMARY

Deliver America proposes the public acquisition of Amazon’s Delivery Network and its integration with the United States Postal Service, creating a single national network under a federally chartered public authority. The acquisition would be:

On Day One, approximately 1 million workers would convert to federal employment. Operational integration would follow on a deliberate schedule, led by career USPS leadership and Amazon alumni, with congressional oversight.

As automation reduces the labor the network requires, the resulting savings would be redirected through an Automation Dividend to worker transition support, rural service expansion, and lower shipping costs for American businesses and households.

The result would be a postal service carrying the country’s parcels at the 21st Century scale, under the power the Constitution grants Congress in Article I, Section 8.

The full proposal, source citations, and forthcoming companion papers are available at DeliverAmerica.org

1. Introduction

The Post Office is In Trouble

The United States Postal Service (USPS), colloquially “The Post Office,” is losing money. Some Americans don’t think of this as a problem: The Post Office is a public service, they reason, and it shouldn’t have to turn a profit. Many do not realize that the USPS is an independent, self-funding agency. By law, it is funded from its own revenue, and that revenue base is in decline.

The USPS has lost $25 billion over 3 years and $109 billion since 2007. It has been on the Government Accountability Office’s (GAO’s) High Risk List continuously since 2009, and as of 2026, has reached its federal borrowing limit set in 39 U.S.C. § 2005, which caps total Postal Service debt at $15 billion and has not been raised since 1992. Postmaster General David Steiner has reported that if USPS were required to make all mandated payments toward its unfunded liabilities, it could exhaust its cash reserves as early as 2031.

Declining Revenue

USPS First-class mail peaked at 103.65 billion pieces in 2001 and has since fallen by more than 57% to approximately 44.3 billion pieces. Total mail volume peaked at 213 billion pieces in 2006 and has fallen by roughly 47% since. Because postal delivery is largely a fixed-cost operation, that decline has not produced proportional reduction in expenses. Package volume has partially filled the revenue gap, but in FY2025 package volume itself declined 5.7%, with revenue growth coming entirely from price increases.

Lingering Debt

The Postal Accountability and Enhancement Act of 2006 required the Postal Service to pre-fund retiree health benefits at approximately $5.5 billion per year. The Postal Service Reform Act of 2022 repealed the pre-funding schedule, canceled past-due payments, and integrated postal retiree health coverage with Medicare. That reform is done.

The liabilities accumulated while the schedule was in force remain on the books, however, and the amortization payments against them continue to drive the majority of reported annual losses. More than 80 percent of the fiscal year 2025 net loss is attributed to factors outside management’s control. The loss the Postal Service earned on operations, its controllable loss, was $2.7 billion.

The Rise of Amazon Logistics

Meanwhile, other delivery interests have grown to dominate the market. From 2018 to 2024, Amazon’s package volume increased by 688%. In 2024 alone, Amazon signed 31 million square feet of new leases, surpassing its pre-pandemic average. In under a decade, Amazon has become a national delivery service provider on par with FedEx, UPS, and USPS itself. In 2025, Amazon surpassed USPS in parcel volume. Amazon Logistics carried 6.9 billion parcels in the United States in 2025 against the Postal Service’s 6.2 billion, making it the largest parcel carrier in the country.

Amazon is Filling The Gap

Amazon is a quintessentially American success. It is the second largest private employer in the United States, and the second largest company in the world. Few other organizations have had such a profound impact on 21st century commerce, and its influence grows. Americans don’t often send letters anymore. We do, however, order a lot of packages from the internet.

Two Systems Doing One Job

In 2024, Amazon spent $194.3 billion on shipping and fulfillment combined: $95.8 billion on shipping and $98.5 billion on fulfillment operations. That figure represents private capital deployed to replicate logistics capacity the USPS theoretically should serve.

As of 2025, Amazon was the largest single customer of the USPS, generating more than $6 billion in revenue and accounting for roughly 25% of all packages the agency delivered, and between 30% and 40% of packages in some rural areas.

That relationship is now contracting. In April of 2026, the two parties revised their agreement, reducing Amazon’s USPS volume by 20%. Layered onto cumulative losses of $25 billion over three years and a cash position the Postmaster General has described as insufficient, this reduction threatens further damage to USPS solvency. In May of 2026, Amazon announced the launch of Amazon Supply Chain Services (ASCS), opening its freight, distribution, fulfillment, and parcel shipping capabilities to businesses of all types and sizes.

In short, the 700 million parcel gap between the two carriers does not represent 700 million new packages. It represents a transfer of carriage from an old network to a new one.

Why This is a Problem

Rural America and the Limits of Private Service

The USPS stands alone in its legal obligation to serve every address in the country (approximately 168 million delivery points) six days a week, at uniform pricing. Federal law prohibits closing post offices solely on the basis of operating at a deficit. This limits cost-cutting options precisely where delivery is most expensive.

What makes this obligation burdensome is not the eye-watering expense of rural delivery, but the collapse of revenue streams that once subsidized it. The USPS absorbs the full cost of rural delivery without the pricing flexibility private carriers are free to exercise. For example: UPS charges Delivery Area Surcharges of up to $19.30 for rural and extended zones atop $5 residential surcharges; their rural pickups can exceed $20. USPS charges nothing extra. The cost of service remains the same for all Americans, no matter where we live.

In 2025 Amazon announced a $4 billion rural expansion to more than 200 new delivery stations targeting 4,000 communities. For Amazon, this expansion is economically viable because it controls both the retail platform through which orders are generated, and the logistics infrastructure that fulfills them. This gives Amazon the ability to anticipate rural routes it can fulfill on its own and build out its network accordingly with no other external shippers involved.

It’s an incredible feat. Private players stepping up to fill gaps in public service could be defended as a market success. It also cannot be repeated by anyone else. Amazon already knows what is going on that truck. It owns the store where the orders are placed, the warehouse the orders ship from, and the van that carries them. It can count the demand its own retail platform generates, and build to that number.

What Amazon is not obliged to make, however, is a public commitment. Their private infrastructure operates where it creates shareholder value.

Price Discipline

According to a class action now certified to represent approximately 288 million consumers, Amazon’s reach allows it to distort pricing across the entire e-commerce market. The core allegation in De Coster et al. v. Amazon.com, Inc. is that Amazon penalizes third-party sellers who offer lower prices on competing platforms, using fulfillment access as a lever to enforce price floors across the internet economy. A seller who cannot afford to lose Amazon logistics access cannot afford to price competitively elsewhere. The consumer, of course, pays the difference.

Privacy and Data Sovereignty

Amazon’s delivery, retail, devices, and cloud businesses combine to observe far more than a shipping address. Opening someone’s mail is a federal crime once it enters USPS custody, a restriction that exists before any postal employee touches your stuff. No comparable federal statute applies once a package enters a private carrier’s hands, including Amazon’s delivery network. What stops Amazon from opening a package, or sharing information about a customer’s mail with a third party? Nothing does, except Amazon’s own word.

The Upshot

Each of these problems compounds the others. USPS is obligated to serve every address at one price; Amazon serves only the routes it chooses to build, priced however it likes. Sellers on Amazon’s platform already pay a premium for the privilege of staying visible, a cost every one of their customers eventually absorbs. And Amazon’s trucks, warehouses, and devices already hold more information about American households than the Postal Service has ever been legally permitted to collect.

None of this changes because Amazon builds more delivery stations or ships more packages. It changes only if the network answers to public law instead of private policy.

What We Should Do About It

Deliver America proposes the public acquisition of Amazon’s Delivery Network and its integration with the United States Postal Service, creating a single national network under a federally chartered public authority.

The acquisition would be:

That last one comes with a catch: nobody outside Amazon knows what this network actually costs to run. Amazon does not report it, and under the accounting rules it does not have to. So the first thing this proposal asks Congress to do is make Amazon open the books. Everything else depends on that number.

On Day One, approximately 1 million workers would convert to federal employment. User data would become subject to USPS protections. Operational integration would follow on a deliberate schedule, led by experienced postal veterans and Amazon alumni, with congressional oversight.

As this network automates, the resulting savings would be redirected through an Automation Dividend to worker transition support, rural service expansion, and lower shipping costs for American businesses and households.

The result would be a postal service carrying the country’s parcels at the 21st Century scale, under the power the Constitution grants Congress in Article I, Section 8.

The Time To Act Is Now

Amazon has already begun opening its logistics network to third-party shippers who do not sell on its platform, reinforcing its function as an independent logistics operator rather than an extension of its retail business. A logistics business with hundreds of customers is easier to acquire than one with thousands. The acquisition case will not improve with time.

The Political Outlook

Deliver America has the potential for broad, bipartisan support.

Infrastructure. Public investments in physical systems that connect the country historically enjoy broad popular support, from canals and bridges to rural electrification and the interstate highway system. Recent polling confirms that sentiment endures: 65% of Americans supported the 2021 Bipartisan Infrastructure Law, including majorities of independents and significant minorities of Republicans.

Economic Fairness. Amazon is the second-largest private employer in the United States. Its 750,000 warehouse workers and 275,000 delivery drivers do not have what post office workers enjoy: federal wage floors, civil service protections, and job security. Most Americans, regardless of party, believe that people doing physically demanding, economically essential work deserve basic protections. Deliver America proposes to extend those protections to a million people on Day One as a direct consequence of public acquisition.

Automation. A January 2025 survey found 89% of U.S. workers are concerned about AI’s impact on their job security. Deliver America proposes that realized labor cost savings from automation be returned to workers and communities. The Automation Dividend is a concrete plan to address structural workforce changes. When automation happens inside public infrastructure, the proceeds should belong to the public.

Privacy. Data privacy is one of the few subjects that still unite Americans across party lines. 72% say there should be more government regulation of what companies can do with personal data, including 78% of Democrats and 68% of Republicans. Amazon’s delivery network sits alongside its retail platform, its device business, and its cloud infrastructure, a combination no comparable carrier holds, restrained by nothing firmer than the company’s internal policies. Deliver America proposes to bring that network under the same public statutory framework that already governs the rest of the postal system, where access to personal information is restricted by law rather than a company’s word. When a network reaches every American address, the rules governing what happens to that information should be public, not proprietary.

How To Take Action

Engage The Framework

This proposal, its assumptions, and models are provided with the hope that they will be examined in public by subject matter experts. If you have ever worked for the USPS or Amazon we are particularly interested in your perspective.

Ask For The Study

The Government Accountability Office is the federal agency that audits things for Congress, and a short bill directing GAO to examine Amazon’s delivery operation and this proposal’s assumptions would settle in a year what no outside analyst can settle at all. That bill is the smallest thing you can ask your representative for and the most useful.

Call your representative’s district office and ask them to look at it. District staff count calls. A university or foundation could also fund an independent version, and that would be worth doing, but subpoena power is the whole point.

Begin The Conversation

Share this proposal with people who make decisions: your representatives, your local officials, journalists, policy researchers. Share it with your friends, family and coworkers. The more people who engage with it seriously, the better it gets.

2. The Constitutional Mandate

The Founders’ Intent

The Founders built the postal system before most other federal institutions because they understood it would be indispensable for democratic self-government. A nation cannot survive absent the means to communicate with and supply itself on its own terms.

Pre-Constitutional Roots

Colonial leadership anticipated that secure communication would be critical to winning independence. Three months after Lexington and Concord, the Continental Congress appointed Benjamin Franklin as the first Postmaster General on July 26, 1775. During the Revolutionary War, the postal system was the connective tissue between army commanders and the first elected representatives, and between representatives and their constituents across hundreds of miles of rough roads.

The Articles of Confederation, ratified in 1781, formally granted Congress “the sole and exclusive right and power of...establishing or regulating post offices from one State to another.” By the time of the Constitutional Convention in 1787, the postal system had already existed in some form for 12 years.

Article I, Section 8, Clause 7

Article I, Section 8 enumerates 18 specific Powers of Congress. Clause 7 reads, “The Congress shall have Power...To establish Post Offices and post Roads.”

Considered alongside the other 17 powers (taxes, borrowing, interstate and foreign commerce, immigration, bankruptcy, currency, counterfeiting, and roads, patents and copyrights, courts, piracy, declarations of war, armies, navies, military governance, and the governance of the capital district), the postal service is a remarkable inclusion. The Postal Service is the only domestic civilian infrastructure function in the list. Everything else concerns national revenue, defense, commerce, or the courts.

Inclusion of The Post Office alongside war powers and monetary authority is a considered statement about the significance of delivery infrastructure.

The Necessary and Proper Clause

Article I, Section 8, Clause 18 grants Congress the power “To make all Laws which shall be necessary and proper for carrying into Execution the foregoing Powers.” The postal clause’s directive to “establish” post offices and post roads becomes, through the Necessary and Proper Clause, a broad legislative grant to do whatever is required to fulfill that function. While the primary infrastructure of American commerce has shifted from post roads to fulfillment networks, the constitutional mandate to maintain that infrastructure has not expired.

The Postal Act of 1792

Washington did not wait for the mandate to prove itself. In his first annual address to Congress on January 8, 1790, not yet a year into his presidency, he named the post office and post roads in the same breath as national defense, naturalization, and a uniform currency, urging Congress toward what he called the intercourse between the distant parts of our Country.

Congress answered two years later. The Postal Act of 1792, signed February 20, 1792, priced a newspaper at 1 cent within 100 miles and 1.5 cents beyond, against 6 to 25 cents for a letter of the same distance, and let printers exchange copies with one another free of charge. It made it a criminal offense for a postal officer to open mail not addressed to them. And it placed the authority to designate new post routes in Congress rather than the Postmaster General, so the network’s reach would answer to elected representatives rather than to whichever route looked cheapest to run that year.

None of this was an accident of drafting. A newspaper cost a fraction of a letter’s postage because the Congress that wrote the rate schedule treated an informed public as a return worth subsidizing on its own terms. When circulation still lagged in the years that followed, Washington pressed Congress to look at the newspaper rate again, on the reasoning that facilitating the circulation of political intelligence deserved a remedy if the numbers showed the current rate wasn’t working. The Postal Act of 1792 is where the constitutional mandate stopped being a clause and became an institution, and the institution was built to move information as deliberately as it was built to move mail.

The Postal Reorganization Act of 1970

By the late 1960s the Post Office Department was straining under 80 billion pieces of mail a year, moved through aging facilities against a growing budget deficit. Postal workers, paid too little to keep pace, staged the first nationwide postal strike in US history in March 1970. Congress's response used its existing Article I postal authority to reassign who exercised it, moving day-to-day control from direct congressional administration into an appointed board.

President Nixon signed the Postal Reorganization Act into law on August 12, 1970 (Public Law 91-375), effective July 1, 1971. The act abolished the Post Office Department as a Cabinet-level executive department and re-established it as the United States Postal Service, an independent establishment of the executive branch. It ended the patronage system under which postmasters were appointed through congressional endorsement, moved rate-setting to an independent process, and phased the agency off annual congressional appropriations toward operating on its own revenue.

Governance authority moved to a Board of Governors of 11 members. Of these, 9 are appointed by the President and confirmed by the Senate to staggered 9-year terms, with no more than 5 from any one political party. Those 9 select the Postmaster General as the board’s 10th voting member; the Postmaster General and the other 9 governors together select the Deputy Postmaster General as the 11th. The House approved the conference report by a vote of 338 to 29.

This proposal would build directly on that governance structure rather than create a new one. The Deliver America Authority would be an expansion of the board Congress established in 1970: additional governors and officers, appointed under the same statutory process, tasked specifically with the merger for its first 3 to 5 years, operating inside the independent establishment the Postal Reorganization Act already created.

Historical Precedents for Public Acquisition of Private Infrastructure

Kohl v. United States (1875)

The Post Office is so vital to the function of the United States government that the very first federal eminent domain case was decided over postal expansion. In Kohl v. United States, 91 U.S. 367 (1875), the Supreme Court held that the federal government possesses the inherent right to acquire private property for public use within the states, as necessary to fulfill its constitutional functions (in this case the establishment of post offices). The Court declared that this authority "is essential to a nation's independent existence and perpetuity."

Deliver America does not propose the acquisition of any assets via eminent domain. Kohl v. United States is relevant, however, because it demonstrates the power to acquire infrastructure in service of the postal mandate is not a new idea. Adequately provisioning the post office is essential to government function, recognized by the Supreme Court 150 years ago, and rooted in the same Article I authority that established The Post Office in the first place.

The Tennessee Valley Authority (TVA)

President Roosevelt signed the Tennessee Valley Authority Act on May 18, 1933, creating the Tennessee Valley Authority (TVA) as a federally owned corporation. It is not-for-profit and has funded all of its operations since 1999 from power sales rather than federal appropriations. It raises capital through bonds sold in the public markets without a federal guarantee. A 9-member board nominated by the President and confirmed by the Senate chooses its chief executive and sets rates, which by statute must stay as low as feasible while covering costs and debt service. It is the largest public power provider in the country, serving roughly 10 million people at rates below those most large United States utilities charge.

The Rural Electrification Act

In 1936, nearly 90 percent of farms lacked electric power because the cost of transmitting electricity to rural areas was prohibitive. Private utilities had studied the problem and declined to solve it; the math did not work for shareholders. The Rural Electrification Act was signed into law by President Roosevelt on May 20, 1936, providing federal loans for the installation of electrical distribution systems to serve isolated rural areas, channeled through cooperative electric power companies. By 1950, close to 80 percent of U.S. farms had electric service. The Rural Electrification Act's core logic, that markets will not serve dispersed populations at uniform prices without public intervention, applies with equal force to rural parcel delivery in 2026.

The Interstate Highway System

Before 1956, the United States had no connected national network of high-speed roads. Limited-access highways existed in scattered segments, most built by individual states through toll authorities that issued bonds and charged drivers to repay them. The Pennsylvania Turnpike, opened in 1940, was the first and prompted similar projects in other states, but these were isolated routes serving single corridors rather than a system. A national network required financing and uniform standards that no single state could supply on its own.

The Federal-Aid Highway Act of 1956 provided both. It authorized 41,000 miles of the Interstate Highway System built to common engineering standards, with the federal government paying 90 percent of construction costs through a dedicated Highway Trust Fund financed by motor fuel taxes. The original authorization was $25 billion, equivalent to over $200 billion in today's dollars. The states retained operation of the roads; the federal role was to fund the network and set the standards that kept it continuous across state lines.

Where a state toll road already followed an Interstate route, the Act designated it into the system rather than buying it. The network was financed by the public and built to a common standard, operated throughout by non-federal authorities.

The Transportation Security Administration (TSA)

At the time of the September 11 attacks, myriad private security companies managed air travel security under contract to individual airlines. Congress created the Transportation Security Administration through the Aviation and Transportation Security Act, signed into law on November 19, 2001, setting a one-year deadline for federal officers to take over passenger and baggage screening nationwide. TSA met that deadline, replacing ~16,500 private airport screeners with more than 40,000 federal screeners across 429 commercial airports by November 19, 2002. This replacement of a private, contractor-based workforce with a federal one, at this scale and on this timetable, is the most direct operational precedent for what Deliver America proposes.

These precedents establish a consistent principle: when private infrastructure becomes essential, the public should have a stake in it.

Couldn’t We Just Regulate Amazon?

The Hepburn Act

The Hepburn Act of 1906 brought railroads and oil pipelines under federal regulation as common carriers required to carry all shippers on equal terms. Standard Oil had used its control of the interstate pipeline network not merely to transport oil, but to determine who else could participate in the oil business. Congress responded by obligating them to carry all shippers on equal terms at published, non-discriminatory rates. Standard Oil resisted even after the statute passed, arguing that they could not be common carriers because they had already purchased the oil they carried. The Supreme Court finally rejected this argument in 1914, eight years after the Hepburn Act became law. Whoever controls the conduit controls who can compete, and regulation of the conduit protects the market. This principle should still hold in the 21st century.

Multi-platform integration as an anticompetitive workaround has been a concern, especially with respect to Amazon, for almost a decade. Lina Khan (FTC chair, 2021 to 2025) anticipated the dynamic in 2017. Among her recommended potential remedies was applying common carrier obligations and duties to Amazon directly, the same legal category the Hepburn Act imposed on pipelines. It’s not a bad idea, but we didn’t do it then and much has changed in the intervening 9 years.

Why should the United States trouble itself to acquire the Delivery Network?

  1. Amazon’s logistics buildout has surpassed the USPS and threatens to supplant it. By itself, this is a reason to pursue regulation, but it does not address the ongoing threats to USPS solvency, which would persist.
  2. Bringing Amazon’s Delivery Network into the USPS would protect the privacy of citizen (i.e customer) data; maintaining high federal safety standards to which Amazon is not beholden and which a common carrier regulation would not address on its own.
  3. Working conditions for Amazon employees and contractors are suboptimal (an issue we take up in Chapter 4). Ordinary federal employment would normally be a step down here: federal agencies enforce workplace safety internally, without the OSHA citations and penalties that already apply to Amazon as a private employer. The Postal Employees Safety Enhancement Act of 1998 is the exception. It wrote USPS into the Occupational Safety and Health Act as an employer, subject to the same enforcement that covers Amazon’s workforce today. It would extend automatically to the Delivery Network’s workforce once absorbed into USPS, without new legislative language required beyond the acquisition itself.
  4. Common carrier regulation would leave the automation trajectory entirely inside Amazon's warehouses and vehicles, with no claim on savings for the workers those systems replace. The Automation Dividend would direct a defined share of automation's labor-cost savings toward worker transition, rural service expansion, and bond repayment instead.

Common carrier legislation would likely take years and split along party lines, and Amazon’s position would harden meanwhile. A negotiated acquisition, financed by bonds offered to the public, would resolve the question as a market transaction rather than a prolonged legal fight - the Hepburn Act’s own pipeline provision took eight years to survive challenge.

Why Not Privatize the Postal Service Instead?

If Amazon's logistics network already outperforms USPS, why not go the other direction and let Amazon run it?

  1. Universal service depends on a legal obligation, not operator efficiency. USPS must deliver to every address in the country six days a week at one price. Amazon already forwards unprofitable rural deliveries to USPS today; owning the whole system would remove the entity currently absorbing those routes, with no substitute.
  2. Rate discipline would weaken. USPS pricing on products like First-Class mail is capped by the Postal Regulatory Commission and tied to inflation, the market-dominant rate ceiling under 39 U.S.C. § 3622. A privatized system might answer to shareholders rather than a rate regulator, particularly in regions with no competing carrier at all.
  3. Mail privacy protections tied to government custody might not transfer. The warrant requirement for opening mail rests on Fourth Amendment protections against government search, dating to Ex parte Jackson in 1878. A private operator would answer to its own policy instead - maybe. Amazon has faced two federal enforcement actions over privacy violations in the past three years (FTC v. Ring LLC and United States v. Amazon.com).
  4. Election mail and official government correspondence would sit inside a single, politically and economically powerful private company, rather than an agency answerable to Congress.
  5. Postal workers would lose the civil service status and collective bargaining protections that do not currently extend to most of Amazon's own workforce, detailed further in Chapter 4.
  6. Privatization would concentrate the country’s only universal delivery network inside the company that already dominates e-commerce fulfillment.

Privatization of USPS would erode basic protections that Americans have relied upon for centuries. Deliver America proposes the opposite direction: public custody of critical infrastructure.

3. Proposed Acquisitions

Defining the Transaction

Deliver America proposes to acquire a specific, bounded set of assets: the network through which Amazon.com, Inc. moves a package from warehouse shelf to doorstep. This proposal names that network the Delivery Network. It would comprise:

It would not include the retail marketplace, Amazon Web Services (AWS), the advertising business, the devices and entertainment divisions, or the Prime membership program, all of which would remain with Amazon. The precise boundary within Amazon’s logistics operation, including which adjacent services would stay with Amazon, is drawn in the Perimeter section below.

The Delivery Network, once acquired, would be integrated with the United States Postal Service and governed by an expanded Board of Governors, the body Congress created in the Postal Reorganization Act of 1970 and the same board that already selects the Postmaster General and sets rates for the Postal Service today. Additional governors and officers, appointed under the same statutory process, would be tasked specifically with the merger for its first 3 to 5 years. We call this expanded board the Deliver America Authority. It would remain part of the Postal Service, keeping the acquisition inside the financing and rate structure Title 39 already provides.

One obvious objection is that the Amazon Delivery Network as defined is not a standalone company waiting to be bought. Amazon does not report these departments as a separate financial segment, and their operations are woven into the retail business at the level of inventory forecasting, demand planning, and order routing. Their separation will require work. This paper models the effort at the level public data allows.

Three facts make this work achievable:

  1. Amazon facilities, fleet, and delivery workforce are managed separately from the retail and technology divisions.
  2. Amazon already sells logistics capacity to outside parties through Fulfillment by Amazon and, more recently, by opening its delivery network to third-party shippers who do not sell on Amazon at all.
  3. Integration exists at the software and data layer. Warehouses, trucks, and sortation equipment transfer cleanly; the systems connecting them to Amazon’s retail operation may be addressed through licensing agreements that define the boundary.

Asset Inventory

The figures here are drawn from public filings and from third-party network tracking, principally MWPVL International. MWPVL establishes the scale and composition of the network with reasonable confidence. These figures are not a substitute for the audited carve-out financials this transaction would require.

Facilities

Within the United States, Amazon operates more than 600 logistics facilities of all types, including fulfillment centers, sortation centers, delivery stations, inbound cross-docks, and air gateways, totaling roughly 435 million square feet, per MWPVL International first-quarter 2025 data. The global network is larger, at approximately 1,200 facilities. The transaction contemplated here concerns the United States network. That count predates a clean split by facility type, and some buildings in it are Amazon Warehousing and Distribution centers that would sit outside the perimeter drawn below. Appendix A flags the facility type that a valuation would require.

One valuation question cannot be answered from public data: the division between owned and leased facilities. Amazon leases a substantial share of its operational space. Leased facilities are not owned assets that pass by sale; they must be assigned or renegotiated, each subject to its own terms and to landlord consent. Establishing the owned-versus-leased split, and the net book value of the owned portion, requires Amazon’s property records and is a precondition to any credible valuation.

Fleet

Amazon’s branded transportation fleet includes approximately 40,000 semi-trucks, 30,000 delivery vans, and approximately 110 aircraft operated through Amazon Air, together with more than 80,000 trailers and more than 20,000 intermodal containers. These are global figures; the United States accounts for the majority but not the whole. Three distinctions bear on acquisition:

  1. The 30,000 branded vans are a fraction of the vehicles delivering Amazon packages, because most last-mile delivery is performed by Delivery Service Partners (DSPs) using vehicles those contractors own or lease, which do not transfer with the network
  2. The aircraft and a portion of the trailer fleet are leased rather than owned, so here again the figures state operational capacity, with the owned share awaiting financial confirmation.
  3. The line-haul trucks and trailers that carry parcels through the network would be acquired, while assets dedicated to the commercial freight service Amazon sells to industrial shippers, including intermodal and rail operations, would remain Amazon's under the perimeter drawn below. The two share vehicles and lanes, so which specific assets would transfer is a question for an audited carve-out.

Automation

More than 1 million robots operate across the fulfillment network, a threshold Amazon reported crossing in 2025, with robotic systems now involved in roughly 75 percent of delivery volume. These systems are developed and produced by Amazon Robotics, the wholly owned subsidiary that began as Kiva Systems, acquired by Amazon in 2012. Amazon Robotics holds the patents, designs, and production capacity behind the deployed fleet. Its intellectual property, more than the unit count, is the consequential asset, because it informs whether the fleet would become dependent on its former owner for replacement and upgrade. The treatment of that intellectual property would be a defined term of the acquisition agreement.

Operational Technology

Routing algorithms, network-planning systems, warehouse-management software, and fleet-coordination tools direct the physical assets above. This layer is the most deeply integrated with Amazon's retail and inventory platforms and the least amenable to clean transfer. Its disposition, by transfer or by license, would also be a defined term of the agreement.

Workforce

The network is operated by more than 1 million workers. The composition of that workforce, the share directly employed against the share contracted, and the obligations attaching to each, are the subject of Chapter 4.

What Public Data Cannot Establish

Three inputs essential to valuation lie outside public reach:

  1. The owned-versus-leased division of facilities and fleet
  2. The net book value of owned property and equipment
  3. The capitalized cost and depreciation schedule of the robotics fleet.

The valuation that follows proceeds on public estimates and states our assumptions.

The Perimeter: What Would Be Acquired, and What Would Not

Amazon today markets its logistics capabilities as a single integrated platform. Amazon Supply Chain Services (ASCS) offers sellers an end-to-end chain: pickup at manufacturing facilities abroad, cross-border ocean and air freight with customs clearance through Amazon Global Logistics (AGL) and Seller Export and Delivery (SEND), bulk storage through Amazon Warehousing and Distribution (AWD), bulk movement to any sales channel through Multi-Channel Distribution (MCD), order fulfillment through Fulfillment by Amazon (FBA) and Multi-Channel Fulfillment (MCF), and parcel carriage to the customer's address, including through Amazon Shipping, the carrier service Amazon sells to outside shippers.

A serious description of an acquisition must state which of these layers would pass to public ownership.

The proposed list of acquisitions follows from the Constitution. Article I, Section 8, Clause 7 gives Congress the power to establish "Post Offices and post Roads," the carrying of items to addresses across the country that the Postal Service exists to perform. The Delivery Network, as this proposal defines it, would be the part of Amazon that now does that same job: the fulfillment centers where orders are picked and packed, the sortation centers and the long-distance trucks and cargo planes that move packages between facilities, the neighborhood delivery stations and the people who carry packages to the door, and the software and robots that run all of it. The government would buy that operation, its buildings, equipment, and workers, but not the goods passing through it, which would still belong to the sellers and shippers who own them. Enabling legislation could apply a simple test: an asset or service would be inside the acquisition if its job is moving goods to American addresses, including the holding of goods that filling an order requires. It would be outside if its job is renting out storage as a service, managing supply chains, or shipping across borders.

Under that test, Amazon would keep its entire supply-chain services business and go on selling it. That includes its storage business (Amazon Warehousing and Distribution), which rents out bulk space for long stretches; its bulk-distribution business (Multi-Channel Distribution), which moves inventory in volume to wholesalers and stores; its international shipping and customs operations (Amazon Global Logistics and Seller Export and Delivery); the Partnered Carrier Program, which is a way of buying trucking rather than a thing to own; Buy with Prime, a checkout button Amazon places on other companies' websites; and the software that coordinates all of these. Every one of them would keep operating, and every one would become a customer of the public network.

What would change is the last step. Goods headed to American customers would be handed to the public network to fulfill and deliver: for Amazon's own packages, under a long-term contract; for every other shipper, at posted rates that are the same for everyone.

The distinction between storage incident to carriage and storage as a product is important, because the two occupy similar buildings. A fulfillment center holds inventory briefly and for one purpose: so that an order placed today can reach an address tomorrow. That holding is part of carriage, as surely as a package resting overnight in a postal processing plant is part of mail delivery. A bulk storage contract, by contrast, is a warehousing service, a business the Postal Service has never been in and should not enter. The enabling legislation could fix the mission boundary in statute: the network would exist to deliver, and the adjacent businesses of warehousing, supply-chain management, and international freight would remain where they are today, in private hands, as customers of the public network.

In short: Deliver America would acquire what duplicates and supplants the Postal Service, and nothing beyond it.

Valuation Methodology

How much is Amazon actually spending (and earning) on delivery? Let’s start with numbers we can trust. In 2025, Amazon’s delivery operation carried about 6.9 billion packages within the United States, more than the Postal Service’s 6.2 billion, more than United Parcel Service, more than FedEx. That count comes from the Pitney Bowes U.S. Parcel Shipping Index, the standard industry reference for measuring package volume by carrier. Measured at the per-package rates Pitney Bowes uses to compare carriers against one another, moving those 6.9 billion packages represents roughly $33.4 billion a year of delivery work, or about $4.84 per package.

That per-package figure is interesting. It is much lower than what established carriers charge for the same kind of work: the index puts United Parcel Service near $15.86 per package and FedEx near $17.04. Part of the gap is ordinary economics. Amazon’s packages are small and light, averaging 1.8 pounds against more than 8 pounds for the larger carriers, and lighter packages cost less to move.

The last mile, though, is only part of what this proposal would acquire. The rest is fulfillment: the warehouses where goods are picked from shelves, packed, and staged for carriers. Here the public record goes dark. Every public company must file an annual report with federal regulators that breaks out results by the divisions management uses to run the business. Amazon runs three: North America, International, and its cloud computing business. Surprisingly, even though Amazon is now the largest shipping concern in the nation, delivery and fulfillment is not listed as a division. It is treated the way a supermarket treats its stockroom: an internal cost of selling things.

Anyone trying to estimate the scale of Amazon logistics will be confronted with 3 gaps in Amazon’s SEC filing:

No outside analyst can separate the network from the rest of the company using what Amazon currently publishes.

A reasonable next question is whether Amazon’s published prices can fill the gap? They can’t, and the reason is instructive. Amazon publishes rate cards for two fulfillment services: Fulfillment by Amazon, for sellers on Amazon’s own marketplace, and Multi-Channel Fulfillment, the identical warehouse work for orders placed anywhere else. For the same box, the same weight, the same pick from the same shelf, the second service costs roughly double the first.

Neither is really the price of fulfillment. The cheaper rate is subsidized, because Amazon collects a commission and advertising revenue on marketplace orders. The pricier rate is defensive, high enough that Amazon doesn’t become a discount shipper for its own competitors. As of 2026 marketplace rates vary further still, pegged to the retail price of the item in the box. Of course, this is only possible because Amazon owns the marketplace and can see what every item sold for.

If USPS splits this network off from Amazon, it would keep the warehouses and the trucks, but not the online store. It couldn’t see an item’s retail price and collect a cut for carriage. By law, it would have to charge one open rate to every customer and Amazon’s rates don’t work that way. They’re priced around keeping sellers loyal to the marketplace, not what it actually costs to move a package. If the United States wishes to purchase this critical infrastructure, how do we estimate its value? Can it pay for itself, the way the Postal Service is required to by law?

We can confidently estimate a supportable price, using arithmetic familiar to anyone with a mortgage. A buyer who borrows a fixed sum at a fixed rate over 30 years owes a fixed annual payment. Lenders additionally require that a business’s operating profit exceed that payment by a margin of safety, since a single bad year should not mean default. Whether that profit exists at a given price depends on operating margin, the share of every dollar of revenue that remains after costs.

Delivery is a thin-margin industry across the board. UPS earns roughly 9 cents of operating profit on the dollar. FedEx earns roughly 6. Amazon’s own North America retail segment earns roughly 7. Old Dominion earns 28 and is considered exceptional. The Postal Service currently earns less than zero. UPS and FedEx are the right benchmarks here because they are the two large American companies whose core business is delivering parcels to doorsteps, the same business as the network USPS would acquire. Let’s say a realistic range for a well-run delivery network, then, runs from 6 to 10 percent.

Apply that range to the network’s likely revenue, and the resulting operating profit could support a purchase price of roughly $55 billion to $85 billion, financed over 30 years at a rate that reflects bonds carrying no federal guarantee. Prices meaningfully above that range would require margins no delivery company operating in America actually earns.

A single number would resolve the range into a figure: the network’s true operating margin, calculated for the assets this proposal would acquire. That number exists today only inside Amazon. Compelling its disclosure, through the audit and rulemaking authority Congress holds over the Postal Service’s competitive operations, would be the first task of the enabling legislation described later in this paper.

Why The US Should Not Acquire Retail or Third-Party Seller Services

The Infrastructure Distinction

The United States distinguishes between the routes commerce travels and the commerce itself. The government builds and owns post roads so private merchants may carry their goods along them; it dredges and maintains rivers so private freight can move on water; it finances the Interstate Highway System so private trucks and private stores can conduct their business. In short: the public provides right-of-way, which private parties use to buy and sell. No one asks whether a federal highway makes the government a trucking company.

Deliver America would bring the Delivery Network to the public side of that equation. The government would own the network that moves the goods and would have no part in selling them.

Including the marketplace would place the federal government in direct commercial competition with Walmart, Target, every small retailer, and every rival marketplace in the country, setting prices and stocking shelves against private merchants who have no comparable backing. Deliver America explicitly does not propose this.

Is This Communism?

No.

The charge that public ownership amounts to Communist capture has little purchase against a government that leaves the selling to private interests. Infrastructure in public hands is the ordinary American arrangement: the postal system; the highways; the air traffic control system; the ports, locks and dams, etc. Confined to logistics, the proposed acquisition stays on the side of the line the United States has occupied since its founding.

Amazon.com, Inc. is the defendant in three active antitrust proceedings concerning the same underlying conduct, a federal consumer class action, a state enforcement action, and a federal antitrust enforcement action, all set for trial in 2027.

The first is De Coster v. Amazon.com, Inc. Filed in 2021 and certified on August 6, 2025, it covers an estimated 288 million consumers, which class counsel call the largest class ever certified in the United States. The claim arises under Section 1 of the Sherman Act. Plaintiffs allege Amazon charges third-party sellers fees well above competing marketplaces, then bars them from pricing lower elsewhere, so the cost reaches the consumer.

The court found common proof that Amazon’s policies function as a platform most-favored-nation restraint, barring a seller from undercutting its Amazon price on a competing site. Sellers who price lower elsewhere can lose the access to Buy Box, have shipping suspended, or face removal. Plaintiffs’ economic experts estimated the conduct affected roughly 288 million consumers across 34 billion transactions. The class period runs from May 26, 2017; consumers may opt out until August 31, 2026; trial is set for June 14, 2027, seeking treble damages and injunctive relief.

The second is California v. Amazon, filed by Attorney General Rob Bonta in 2022 in San Francisco Superior Court under state antitrust law. On April 20, 2026, Bonta announced the public release of evidence his office called proof of an illegal price-fixing scheme. The unredacted filing describes communications in which Amazon pressed vendors to raise prices on competing retailers, among them Walmart, Target, and Best Buy. A hearing on the state’s preliminary injunction is set for July 23, 2026, and trial for January 19, 2027.

The third is Federal Trade Commission v. Amazon.com, Inc., filed September 26, 2023 and joined by 17 states. Before the same judge presiding over De Coster, it alleges Amazon illegally maintains monopoly power through two mechanisms: an anti-discounting scheme that punishes sellers for pricing lower elsewhere (the same conduct at issue in the other two proceedings), and a tying arrangement that conditions favorable product placement on a seller’s use of Amazon’s own fulfillment and delivery network. A federal judge ruled in October 2024 that the case may proceed on its central allegations. Trial is scheduled for March 29, 2027. The FTC seeks a permanent injunction and other equitable relief rather than damages.

Structural Relief

Going forward, the Deliver America acquisition could remove the fulfillment-access lever that gives both California v. Amazon and the FTC’s claim much of their force. That’s what’s known as structural relief, exactly the kind of fix these government lawsuits are asking the courts to order. It could also end Amazon’s legal exposure on this conduct for good, whatever happens with the pending cases.

The disputed policy works through penalties Amazon imposes on a seller who prices lower on a competing platform. Those penalties draw on two separate powers. The first is control of fulfillment: a seller who loses Amazon’s fulfillment access loses the means to deliver, the power the FTC’s tying claim targets directly. The second is control of demand: Amazon’s marketplace holds the dominant share of online marketplace sales, about 72% by the De Coster plaintiffs’ estimate, such that removal from the platform severs a seller from where the buyers are.

Public ownership would affect the first power directly. A network owned by the USPS and operated at regulated, non-discriminatory rates would serve every shipper on equal terms. A seller who ships through it would no longer depend on Amazon to deliver and would no longer forfeit fulfillment by pricing lower elsewhere. The most coercive of Amazon’s levers would be neutralized.

The Anchor Customer

If Deliver America is successful, Amazon retail would become the largest customer of a network it no longer owns. Amazon’s Delivery Network was built to move Amazon’s own goods. If the Delivery Network becomes part of the USPS, Amazon retail’s volume is likely to remain its largest source of demand for the foreseeable future. A customer of that scale and steadiness is a foundation to build on.

A long-term service agreement would commit Amazon retail to the USPS Delivery Network through a transition period, perhaps 5-10 years, with a volume commitment agreeable to both parties. Rates would follow a published schedule that declines with volume and applies to every shipper. Amazon would occupy the most favorable tier because of the volume it moves, not because of preference; any shipper at the same volume would receive an identical rate. Amazon, having handed off the capital and operational weight of running its own network, would keep a mature carrier it can rely on without the cost of sustaining it.

4. Workforce Transformation

Documented Current Conditions

The strongest evidence of warehouse conditions is Amazon's own injury data, reported to the Occupational Safety and Health Administration and analyzed by a Senate committee in December 2024. On the company's reported figures, its warehouses recorded injuries more than 30 percent above the warehousing industry average in 2023, ran at nearly twice the industry rate over the preceding 7 years, and exceeded the average at more than two-thirds of facilities.

The pace that produces those injuries is set and enforced by software. Amazon's system logs the interval a warehouse worker spends not scanning items as Time off Task, and by the company's own account the system can generate a warning or a termination from accumulated Time off Task without a supervisor's review. Amazon has separately patented a wristband designed to track a warehouse worker's hand and arm movements and direct them toward inventory bins in real time. Delivery drivers face a comparable structure: many partner-firm drivers are required to run a monitoring application on their phones that records vehicle location, speed, braking, and, in some documented cases, in-cab video, feeding into performance and disciplinary decisions. Four states now require warehouse employers to disclose production quotas and the discipline attached to them.

A workforce paid at these levels is also on record as a workforce the public is already partly supporting. The Government Accountability Office found Amazon among the larger employers of Supplemental Nutrition Assistance Program and Medicaid recipients in the states it examined, at a time when fewer than 10 percent of aid recipients nationally worked for a public-sector employer of any kind.

These same roles are marketed heavily to one population in particular. Amazon has built a large, public recruiting apparatus around veterans and military spouses, explicitly framing warehouse and driving work as suited to people trained to make fast decisions under pressure. Amazon's own director of military affairs has stated publicly that a military spouse relocates roughly seven times more often over a career than a civilian spouse, and that many military spouses leave the company when a service member receives relocation orders; the company built an internal retention program, Project Juno, in response.

Clearly, Amazon already understands the advantages of a workforce with publicly funded benefits.

The Delivery Workforce

Most last-mile delivery runs through Delivery Service Partners, independent firms operating across the United States whose drivers are their W-2 employees, not Amazon's - one contractual layer removed from the network they serve.

A separate program, Amazon Flex, engages drivers who use their own vehicles to deliver packages in blocks booked through an app. Amazon classifies these drivers as independent contractors. State authorities have rejected that classification, however. The Virginia Employment Commission found that Amazon had misclassified its Flex drivers, who for unemployment-insurance purposes were employees rather than contractors; the Virginia Court of Appeals upheld that finding in 2023, and the Supreme Court of Virginia left it in place in 2025. The Wisconsin Supreme Court reached the same conclusion under that state's law.

The benefits and protections for these workers depend on which entity employs them and on hours worked. Full-time warehouse staff receive Amazon coverage. Drivers employed through partner firms are covered by those firms, so coverage varies. Flex drivers, classified as contractors, receive no employer coverage at all.

None of these workers hold the standing of federal postal employment: enforceable safety standards, a grievance process, collective bargaining, statutory wage floors, and a path to a career position with full benefits. Extending that standing would be the first consequence of acquisition.

The Overnight Transformation

On the day the acquisition closes, the workforce of the Delivery Network that is directly employed by Amazon would become federal employees under the Deliver America Authority. Operations would continue unchanged, with the same workers reporting to the same facilities and running the same routes the following morning. Operational integration would proceed on the schedule set out in Chapter 7.

Career vs. Non-Career

Federal postal employment is structured in two tiers. Career positions carry the full federal package of benefits. Non-career positions carry a more limited set of benefits and a defined path to career status.

The package attached to a career position includes health coverage, enrollment in the federal retirement system, statutory wage floors, a grievance and arbitration process, and collective bargaining. It also includes a protection particular to postal employment: while most federal agencies are covered by an internal safety program, The Postal Employees Safety Enhancement Act of 1998 applies the Occupational Safety and Health Act to the USPS in the same manner as to any private employer. Before that Act, the Postal Service was exempt from those private-sector provisions. Workers now covered by OSHA as private employees would lose no safety standing.

The delivery workforce would reach federal employment by a different route than the directly employed. It is composed of two groups: the larger group works for Delivery Service Partners (DSP), the independent firms that run Amazon's delivery routes under contract; those drivers are the firms' employees, not Amazon's. The smaller group is engaged through Amazon Flex and classified as contractors in most states ( Virginia and Wisconsin authorities rejected this in favor of employee status.)

It is possible that many current Flex drivers occupy that role specifically for its flexibility and do not seek full time government work. For these, and other individual contractors, it is reasonable to offer non-career positions inside the USPS, analogous to seasonal or non-career positions that already exist.

Precedent

There is a 21st century precedent for rapid conversion of a large contract workforce into federal employment. Congress passed the Aviation and Transportation Security Act on November 19, 2001, setting a one-year deadline for federal officers to take over passenger and baggage screening at every commercial airport in the country. The first airport, Baltimore/Washington International, converted to federal screening on April 30, 2002. All 429 commercial airports had converted by the November 19, 2002 deadline, and the resulting federal screener workforce exceeded 40,000 employees, replacing roughly 16,500 private contract screeners.

A single statute converted the function nationwide. The government did not renegotiate with each of the hundreds of private screening companies operating at individual airports; at each airport, on a set date, the private contract ended and a federal screening force took its place. The relevant precedent is, government can quickly convert a function of this kind, by statute rather than by thousands of separate agreements.

The individual employment guarantee is where enabling legislation for Deliver America might differ by design. TSA terminated the private screeners at each airport when their contract ended, then ran individual screeners through a competitive assessment, background investigation, and citizenship check before they could join the federal workforce. Many did not pass or did not reapply, and the government anticipated meaningful attrition in the process, particularly at larger airports. The Department of Labor provided transition assistance to the screeners the conversion left behind.

Deliver America proposes to borrow the mechanism, a single statute converting the function nationwide on a defined schedule, without the competitive rehiring. How the contracted delivery workforce might convert into the USPS two-tier structure, on a defined timetable, is really a question for the USPS. Ideally, it would happen without reopening each underlying contract, but USPS has its own standards which will attempt to address below.

A Special Case

The roughly 4,400 owners of Delivery Service Partner firms who operate the businesses should probably be approached differently from their employees. Many DSPs are small business owners whose only customer is Amazon.

Enabling legislation could offer each owner a choice. A DPS owner could accept compensation for the terminated contract, covering the unexpired term, vehicles, and other assets, and exit the business. Or, an owner could instead apply for a (station-level?) supervisory role in the new structure, since the dispatch, route assignment, hiring, and compliance work a DSP owner already does will still need doing. The mix of outcomes that would follow, with some owners exiting and others converting into USPS jobs, could resemble what happened to private security firms after TSA federalization: some closed out entirely and others found a continuing role through the Screening Partnership Program.

Deliver America is actively soliciting comment from all Amazon employees, contractors and subcontractors. Please get in touch. Your input is vital to this proposal.

Sidebar: It’s worth noting that a meaningful share of Flex drivers are neither US citizens nor permanent residents, which would disqualify them for USPS employment. Expedited Naturalization in exchange for a period of service, as with the US Armed Forces, would make these roles very desirable for aspiring Americans and incentivise a committed corps of drivers. Deliver America does not explicitly make this recommendation as of this draft, but it’s interesting to think about.

The View from Inside USPS

Bringing the Delivery Network's workers into the Postal Service raises two practical questions. How would they get hired, and what would their arrival mean for the postal employees already on the job? The Postal Service's own rules should be a starting point.

Placement of the incumbent workforce

The Postal Service has a standard way of hiring from outside. It posts a job opening, applicants take a test where one is required, qualified applicants are ranked, and a manager selects from the top of that list. This system works well for filling a handful of jobs at a time, but it was never built for the numbers (or novel roles) Deliver America proposes.

So the enabling legislation would take a different route. It would create a one-time permission to move workers inside the acquisition perimeter into postal jobs directly, as the starting staff of the acquired facilities. They would not have to apply for their own jobs or compete against strangers on a ranked list. Moving an entire working team over along with the buildings and equipment they run is a different situation from picking new hires off the street; law would treat it that way.

The usual safety checks would stay. The Postal Service requires a suitability review, a medical assessment, and in some cases a background investigation before someone starts. Under the conversion, those checks would still happen, but they would run as conditions of keeping the job rather than as hurdles to clear before starting. People could begin right away, and anyone who fails a check would be removed rather than blocked at the door.

Veterans' preference

Federal law gives veterans a hiring preference when they apply for postal jobs from outside, under the Veterans' Preference Act of 1944. The conversion would keep that preference in place for every job that opens in the Delivery Network after the initial move. Once the starting workforce is set, every new opening in the acquired facilities would carry the veterans' preference, exactly as postal jobs do now.

The existing postal workforce

Most frontline postal jobs are filled by seniority. When a position opens, current employees bid on it, and the most senior qualified bidder wins. That bidding right is one of the things postal workers value most, and it’s written into their union contracts. A large batch of new jobs could look, at first glance, like it was skipping past those bidding rights. We want to address that concern directly.

Seniority would work on two separate tracks. For bidding, a converted worker's seniority would start on the day of the conversion. A new arrival could never outbid a current employee for an assignment, and every postal worker on the job today would keep their place in line ahead of every convert. For pay, leave, and retirement, a converted worker's years at Amazon could count, so the promise of full benefits on the first day would still hold.

The law would also promise that the acquisition would cost no current USPS worker their job. It would bring no layoffs, no elimination of routes, and no facility closures aimed at the existing workforce. The Postal Service's layoff procedures would not be used against current employees as a result of the conversion.

Converted workers would join existing craft units and be represented like any other postal employee, with union dues handled through the system already in place. Any jobs above entry level in the new facilities, and every future opening in them, would be filled through the normal bidding process under the union contracts. Over time, the acquired network would fold into the same seniority-and-bidding system the rest of the Postal Service already runs on.

All of this would sit inside the postal labor framework set up by the Postal Reorganization Act of 1970, which includes binding arbitration and a no-strike agreement. How automation savings would be shared, and what they would mean for staffing down the road, would be handled by the separate companion paper on the Automation Dividend.

Set up this way, the conversion would bring the Delivery Network's workers into the Postal Service without taking anything away from the workers already there.

Cost Modeling

Federalizing the workforce would change its cost in two ways: the benefits attached to a federal position, and the effect of statutory wage floors on the lowest-paid roles.

Benefits are the larger factor. The Bureau of Labor Statistics puts benefits at about 30 percent of total compensation in private industry and about 38 percent in state and local government, and a federal package sits at the upper end of that range.

That benchmark is a gross figure, and it overstates the added cost for two reasons. Amazon already pays wages and provides benefits to its directly employed workers, and the partner firms already cover many of their drivers, so the cost that matters is the difference between the federal package and the compensation these workers already receive, not the federal package. The two-tier structure governs the increment: a career position carries the full package, a non-career position less. What federalization adds per worker hinges on where the worker enters and on what the worker is paid today.

For that reason the total cannot be fixed from public data. It depends on the current compensation of a workforce whose carve-out financials do not yet exist, on the split between career and non-career tiers, and on the size of the workforce as automation reduces it over the transition.

Appendix B models the total across conservative, base, and optimistic cases rather than resting on a single number. The cost would fall where any carrier's labor cost falls, inside an operating business that bills its throughput at market rates. The question is whether operating revenue covers labor cost, which Appendix C takes up.

An offset runs the other way. Workers placed on federal wage floors and federal benefits draw less on public assistance, a cost other parts of government now carry. The scale of that offset, and the net after it, belong to the fiscal case in Appendix C.

Pension and Retiree Health Obligations: An Honest Accounting

The Postal Service runs two retirement systems. The Civil Service Retirement System (CSRS), a defined-benefit plan closed to anyone hired after 1983, holds the largest legacy obligations. The Federal Employees Retirement System (FERS) covers everyone hired since. The incoming workforce would enter FERS as new hires.

The legacy problem is pretty serious. The Government Accountability Office reported the Postal Service's unfunded CSRS liability at about $60 billion at the end of fiscal year 2024, and its unfunded FERS liability at about $46 billion, the latter swollen by payments the agency skipped. The Office of Personnel Management sets the pension amortization payments under federal retirement statutes; the agency controls neither the assumptions nor the schedule, and the payment swings sharply year to year. Other federal agencies receive Treasury or appropriated funds for these obligations; the Postal Service must cover them from operating revenue.

The 2022 Postal Service Reform Act repealed the retiree-health prefunding mandate, canceled the past-due prefunding, and folded postal retiree health into Medicare, so that reform is complete. What still requires companion legislation, with or without this acquisition, is CSRS reallocation.

As automation shrinks the workforce, the base paying in would fall while early-cohort obligations accrue. But FERS funds each benefit as it is earned, and the defined-contribution Thrift Savings Plan holds individual accounts that leave nothing collective behind. The Postal Service's current unfunded FERS liability came from skipped payments, and an operating business billing throughput at market rates could make payments the current USPS cannot.

Recommendations

The legacy CSRS liability, including the disputed allocation, should be flagged as a pre-acquisition matter for the Treasury and the Office of Personnel Management. The incoming workforce should enter FERS as funded by law, with contributions kept current. The reallocation that 2022 left unfinished should be completed, and a share of the Automation Dividend (Chapter 5) could be dedicated to retiree obligations. Deliver America would not solve the legacy pension and retiree-health problem, which predates it and needs its own legislation. What it would do is add its workforce to the funded system and bring the operating revenue that lets those obligations be met.

5. The Automation Dividend

When a worker’s task is automated, the labor cost of that task falls. A facility that needs fewer people to move the same volume of packages costs less to run, and the reduction shows up as margin, moving upward to the owners who automated the job away.

If the Distribution Network moves into public hands, the destination of money saved via automation would become a governance decision, made through the board of the Deliver America Authority, rather than a return calculated for Amazon shareholders. We call that savings the Automation Dividend.

The Automation Dividend could potentially fund five things:

  1. Accelerated repayment on the American Commerce Bonds that finance the acquisition
  2. A Worker Transition Fund tied to workers whose specific roles are eliminated
  3. Rural service expansion under the universal service mission
  4. Reduced shipping costs for businesses and households served by the network, and
  5. Continued operational improvement at USPS as the two systems integrate.

For purposes of this paper, we estimate the dividend pool at $20 to 24 billion annually at maturity. This foundations of this estimate, the legislative mechanism that would enact it, and the potential impact of automation on the Delivery Network’s workforce are the subject of an upcoming companion paper in Q3 2026, available at DeliverAmerica.org

6. Acquisition Structure

Deliver America proposes a purchase of assets carved out from the corporate entities that hold them today, transferred to public ownership along with the workforce that runs them. Chapter 3 set out what falls inside that perimeter and what it would be worth.

Deliver America does not propose to buy Amazon.com, Inc., a controlling interest in Amazon, or shares in any company. The target is a functional bundle, carved out from Amazon.com, Inc. and its subsidiaries that hold them today.

The acquisition would be negotiated, and it would be priced at fair market value. Chapter 3 puts that value in a range of roughly 55 to 85 billion dollars, a figure reached from what the network’s likely earnings could support as debt at a rate a real lender would accept, rather than from its replacement cost or a revenue multiple. The proposal does not seek to acquire the network below that value through the pressure of the state. It seeks to pay what the network is worth and to fund the payment as described below.

While the federal government can use eminent domain to acquire private property for public use, (especially for the post office), a negotiated purchase is preferable in every way that matters. A negotiated sale would avoid years of litigation and would keep the network intact and operating through a transition. Further, it would preserve the possibility, developed in the following sections, that the seller has his own reasons to come to the table.

The Transaction Framework

American Commerce Bonds

The negotiated price would be financed, not appropriated. Deliver America proposes to raise the purchase amount through publicly offered infrastructure bonds, sold to institutions and to ordinary citizens, and repaid over time from the revenue the network would generate. We call them American Commerce Bonds.

The country has financed large public undertakings this way before. The Liberty and Victory Loans of the First World War, and the War Bonds of the Second, funded war efforts through direct public subscription, sold in denominations small enough for a household to buy. Municipal revenue bonds build and maintain ports, water systems, and airports across the country every year, serviced by the fees those assets collect rather than by the general tax base. American Commerce Bonds extend that tradition: a public asset, financed by those willing to lend against its revenue, repaid by what it earns.

Bonds would be preferable to direct appropriation for reasons both structural and political. Structurally, they would tie repayment to the asset: there would be no line in the federal budget reading “give the Deliver America Authority billions of dollars.” The debt would be serviced by the network’s operating revenue on a fixed schedule, insulated from appropriations fights. Politically, an appropriation of this size would be difficult to pass and easy to attack. A bond issue is an investment that pays out on a schedule the asset’s earnings can meet, not a tax.

Who purchases the bonds actually matters. A bond holder in the Delivery Network has a stake in the network’s success and real ownership of the outcome. War bonds were sold partly for the money and partly to give millions of people a personal share in a national effort. American Commerce Bonds could do the same: build a constituency for the network’s success at the moment of its financing, distributed across the individuals and institutions who invest.

The structure itself, the term, the interest rate, the repayment schedule, and the mix of institutional and retail offerings, would be set in the enabling legislation and modeled against the revenue scenarios developed in Appendix C.

What's In It for Amazon?

Every shareholder would be paid in full, at a price within the range Chapter 3 establishes, roughly 55 to 85 billion dollars, set through negotiation once the diligence process described there is complete.

For the company itself, the transaction would convert a capital-intensive division into a large cash payment and a long-term service contract. Amazon Logistics carries the physical plant, the vehicle fleet, the injury rates, and the wage and classification litigation that come with employing more than a million people directly and through contractors. A sale would move all of that off Amazon’s books at a fair price, while a long-term anchor customer agreement would preserve the delivery service Amazon’s retail business depends on, without Amazon continuing to own, staff, or defend the infrastructure behind it.

The timing also matters. Amazon faces active antitrust litigation with trial dates across 2027. Selling now, on terms Amazon would help negotiate, would put the company in a materially stronger position.

Capital freed by the sale, and management attention no longer spent operating or defending the logistics division, could be redirected to the businesses generating Amazon’s actual margin: AWS, advertising, and the AI infrastructure buildout the company has prioritized in every recent earnings call. A logistics network built to move one retailer’s parcels is a mature, capital-heavy, comparatively low-margin asset next to those businesses. Selling it at a fair price would be a legitimate capital allocation decision on its own terms.

Sequencing and Transition

The transfer of the directly employed workforce into federal employment would happen at closing, on a single day. The integration of the network with the Postal Service would happen over months.

At closing, ownership of the assets would pass to the public and the workforce would convert to federal employment under the Deliver America Authority as described in Chapter 4. The morning after closing, the same workers would report to the same facilities and run the same routes.

Operational integration would follow on the schedule Chapter 7 develops: selective at first, joining Postal Service and network infrastructure where they complement each other, and fuller over time as the two systems merge. The network would keep running throughout. Nothing about the transfer of ownership would require the network to pause, reorganize, or degrade service while the longer work of integration proceeds.

The word “phase” in this proposal refers to the pace of operational integration, not to a schedule of buying the network in pieces. The Deliver America Authority would take operational control at closing and would hold it permanently. The governance section that follows describes what kind of body it would be.

Governance

The acquired network would be integrated into the United States Postal Service and governed by an expanded Board of Governors, the body the Postal Reorganization Act of 1970 created and the same board that already selects the Postmaster General and sets rates for the Postal Service today. Additional governors and officers, appointed through the same presidential-nomination, Senate-confirmation process that already governs the board, would be tasked specifically with the merger for its first 3 to 5 years. This expanded board is the Deliver America Authority referred to throughout this proposal. It would not be a separate corporation standing outside the Postal Service, and it would not replace the Postal Service.

The Authority would carry the operating weight of the enlarged institution. Its concerns would be concrete ones: how a Postal Service that has absorbed more than a million new workers and taken on service of the American Commerce Bonds would budget and plan against its enlarged revenue, and how the network itself would run from one day to the next. These are management questions, answerable by the people running the network rather than by Congress directly.

Congress would set the Authority’s mandate through the enabling legislation and would retain oversight of whether that mandate is met. What the structure would keep at a distance is day-to-day operational control. Congress would set the mission and review the results. The expanded board, and the Postmaster General who sits on it, would run the network.

The enabling legislation would expand the Board of Governors, authorize the American Commerce Bonds, and convert the workforce to federal employment as a class. A single bill could accomplish the expansion, the financing, and the workforce conversion together, rather than leaving any of the three to a later and less certain vote.

Some of what the expanded board would need is already provided by USPS today. The Postal Service can issue debt to the Treasury at favorable rates, can negotiate volume-based service agreements subject to regulatory review, and since 2022 can offer certain non-postal products and services to raise revenue.

7. The Transition Plan

This chapter is the most speculative of any in the proposal. What follows here are not firm recommendations. Here we flag what we believe might be reasonable milestones and timelines. Analysis and feedback from industry experts will be necessary to craft a transition plan.

Practical Constraints

Ownership transfer of routing software, warehouse management systems, and fleet dispatch tools that have run continuously for a decade carries real risk. Not every facility in the Distribution Network is owned outright. Some operate under lease, and the acquisition would need to sort out which facilities transfer as owned real estate and which require new lease arrangements or renegotiation with third-party landlords. Internal communications would need to scale to accommodate a new department and its personnel. Further complications and "unknown unknowns" will surely surface. A phased schedule will create a roadmap for USPS to absorb that complexity without degrading service quality.

Principles

Continuity of Service is Paramount. The USPS and Amazon are both nationally recognized brands, and user expectations are high. While the USPS is appreciated for its legacy, Amazon has distinguished itself for its ingenuity. Both are known for their reliability. A central goal of this proposal is to maintain those standards throughout the acquisition process. Delays, errors, and customer confusion are unacceptable.

Think Slowly, Act Quickly. Extended timelines create room for uncertainty and dissipation of resources. Deliver America should be approached with the qualified rigor demanded by major public works projects, national security, and economic stability. By the time Congress acts (hopefully sometime between 2028 and 2030), no part of the transition plan should be TBD.

Plan For Success, and Learn From It. This proposal is an invitation to policy, logistics, and technology experts, in cooperation with the public, to build a plan that is executable the day it is ratified. Advances in transportation, automation, communication, and energy will continue to occur alongside acquisition and integration. The Deliver America Authority should include subject matter experts, including rank and file USPS and Amazon workers, who can model the impact of these gains and update the Transition Plan accordingly.

The phased schedule that follows is built to hold to these three principles.

Day One

On the day the acquisition closes, the Distribution Network would pass permanently into public ownership under the Deliver America Authority.

Federal employment status and the benefits package Chapter 4 describes, health coverage, enrollment in the Federal Employees Retirement System (FERS), statutory wage floors, and Occupational Safety and Health Administration (OSHA) enforcement, would take effect immediately for the directly employed workforce.

The DSP transition mechanism, distinct from direct-employee conversion because DSP owners are small business operators rather than employees, would need to be designed and negotiated with the DSP community prior to day one, following the mechanism the Transportation Security Administration's 2001 to 2002 transition established as precedent.

The routing software would keep directing the same trucks along the same routes it directed the day before.

The First 12 to 18 Months: The Continuity Phase

Existing management would remain in place, with articulated career trajectory plans: some would stay, some would leave, and the Deliver America Authority should track those plans accordingly. This acquired layer of leadership would report to the Postmaster General, or to whomever the Postmaster General deputizes inside the Deliver America Authority. Department leadership, appointed under whatever process the enabling legislation settles on, would need to be in place early enough to oversee the following:

The audited carve-out financials Chapter 3 flags as unavailable from public data, if still incomplete, would be confirmed within this time frame, establishing which facilities and fleet the network owns outright, which operate under lease, and the depreciation schedule of the robotics fleet.

Intellectual property licensing agreements will settle which systems transfer outright and which stay licensed from Amazon. The software layer is less straightforward: routing algorithms, warehouse-management systems, and fleet-coordination tools currently feed the same inventory forecasting, demand planning, and order routing systems Amazon retail uses to run its marketplace, so severing them would mean defining exactly where the Delivery Network's software ends and Amazon’s retail and supply chain tools begin, a boundary the licensing agreement would draw. The long-term service agreement with Amazon retail would take effect, locking in the volume commitments and the non-discriminatory rates.

Internal communications systems would need to extend to the USPS’s expanded personnel without creating confusion about who reports to whom during the handoff; the requirements for which should be finalized during this period.

Months 18 to 36: The Integration Phase

A unified technology platform would begin development. Tracking, routing, and customer-facing systems that ran as two separate stacks through the Continuity Phase would merge on a schedule set by technical teams doing the work, rather than one fixed by statute.

Physical rebranding, new signage and updated vehicle liveries, would begin in this window too, rolling out gradually as equipment reaches the end of its service life. A wholesale rebrand that pulls working vehicles and functioning facilities out of service to reskin them is not the plan.

Labor relations would mature in parallel. USPS employees work under decades of collective bargaining relationships with the American Postal Workers Union (APWU) and the National Association of Letter Carriers (NALC). By year three, collective bargaining recognition for the newly converted workforce would move from design into practice, with the two legacy workforces operating under one shared framework. These terms would be negotiated with the APWU and NALC.

The anchor customer relationship with Amazon retail would mature in this window as well, alongside the network's opening to other shippers at published, non-discriminatory rates. Third-party volume would grow from the initial group of shippers who join once the network opens toward the broader market.

Rural service expansion should be measurable by this point: delivery speed and reliability in rural service areas would move toward parity with what customers in dense metropolitan areas already receive, using the routing technology and fleet density both networks bring to the merger.

Year 3 and Beyond: Full Consolidation

By year three, the Distribution Network and USPS would operate as a single network under the Deliver America Authority. The two labor forces that operated in parallel through the first three years would combine into one workforce, and the systems that ran alongside each other during the same period would consolidate into one technology platform. Branding would be the last piece to finish changing.

The Automation Dividend could be realizable during this period. As automation reduces the network's headcount needs against its growing volume, a defined share of the resulting savings would begin flowing to bond repayment, the Worker Transition Fund, and the other allocations on the pace automation occurs.

The long-term service agreement with Amazon retail would continue on the term set at acquisition. Amazon retail would likely remain the network's largest customer, paying the same published rate available to any other shipper of equivalent volume under the non-discrimination provision.

Risk Factors and Mitigation

Not an exhaustive list.

Technology preservation. The Distribution Network's routing algorithms, warehouse management systems, and the institutional knowledge held by the engineers who built and maintain them make up much of what the network is worth acquiring. Retention terms for key technical staff through the Continuity Phase, in addition to well-organized IP licensing, would address this directly: the network would keep access to the people and the systems it needs without a rushed technical handoff.

Culture and operating cadence. Amazon's logistics operation runs on a decision-making speed that federal agencies are not generally built to match, and a mismatch between the two risks slowing the network down. Placing the network inside a self-funding department structured to make independent operating decisions is not sufficient: leadership should draw on subject matter expertise who understand how the network has historically operated.

Political interference. Congressional oversight of a network this size is appropriate. Congressional micromanagement of its daily operating decisions is not. Congress would charter the Deliver America Authority and set its mandate, then review whether that mandate is met; the department itself would run the network day to day.

Labor relations. The incoming workforce and the existing USPS workforce would arrive at this merger from different labor histories. The converting Distribution Network workforce, largely without union representation before conversion, would enter a federal labor relations framework for the first time, while USPS employees would already be operating under established relationships with the APWU and NALC. The risk to manage is timing: running two disconnected labor relations processes longer than necessary would recreate the two-tier structure this merger is meant to eliminate, which is why one shared framework by year three is the target this proposal sets.

Service continuity. A rural customer or an Amazon retail customer should notice nothing about their delivery service changing in year one, and the Amazon retail service agreement's volume commitments would give the network a direct commercial incentive to keep it that way. Integration work carrying any risk of visible service degradation, a merged tracking system or a consolidated sortation facility, for example, would proceed on the schedule the technical work requires rather than one set by statute.

The anchor customer relationship. Amazon retail would be the network's largest customer and its former owner, a position that would carry more commercial leverage than any other customer The long-term service agreement would exist to manage that leverage by setting a defined transition term and volume commitments running in both directions, backed by a non-discrimination provision guaranteeing Amazon retail the same rate available to any shipper of equivalent volume. A gradual opening of Amazon retail's volume to competition from other shippers over the agreement's term would keep this relationship from settling into a permanent dependency in either direction.

8. Political Architecture

The Congressional Pathway

Ideally we could make this happen with a single bill expanding the Board of Governors, authorizing the American Commerce Bonds, and converting the workforce as a class on the TSA model. How do we get there?

The postal reform coalition already exists. The Postal Service Reform Act of 2022 passed with substantial bipartisan majorities, assembled by members who worked the issue for years. Those members, their staffs, and the advocacy organizations that supported them constitute a working coalition that has already passed major postal legislation.

The committee map runs through panels holding postal jurisdiction and oversight, Senate Homeland Security and Governmental Affairs and House Oversight chief among them, with Commerce committees engaged on the carrier and rate questions and Appropriations engaged despite the bond structure, because members of those committees expect to be consulted on anything of this scale whether or not it spends appropriated money. Early staff-level relationships on those committees would matter more than early member endorsements.

One companion measure belongs in or alongside the enabling legislation: the Civil Service Retirement System cost-allocation correction described in Chapters 4 and 8. The Postal Regulatory Commission and the Postal Service Office of Inspector General have both found the current allocation unfair to the Postal Service, at a scale of roughly $75 billion, and the finding has stood for more than a decade without action. A Congress expanding the Board of Governors could settle the old account in the same act.

Natural allies are already seated. Rural-state members of both parties, the labor caucus and Postal advocates in both chambers have spent careers on the institution’s survival. None of these members need persuading that the problem exists but might need persuasion with respect to this proposal. Their insights will likely improve it.

By 2028, we hope to build a coalition, place the proposal before the public, and develop legislative text, so that the idea can be taken up by candidates from any party.

Opposition and Responses

“The government can’t run anything.” The Deliver America Authority would be a self-funding expansion of the Postal Service’s existing Board of Governors, built to keep daily operations apart from the annual political process. It would be run by industry experts, not by Congress. The country has run public enterprises on commercial discipline before; the Tennessee Valley Authority, a federal corporation since 1933, has funded its own operations without a congressional appropriation since 1959, more than 65 years now.

“This is socialism.” The proposal would purchase assets at fair market value, through negotiation, financed by bonds sold to willing lenders and repaid from revenue. No property would be seized, no industry nationalized by force, and no private commerce displaced.

“It would kill Amazon.” Amazon retail, Amazon Web Services, advertising, devices, and every other line of business would be untouched. Amazon would be paid within the range Chapter 3 establishes, roughly 55 to 85 billion dollars, for assets that consume enormous capital to operate, would shed that operating weight, and would retain guaranteed access to a mature carrier under the long-term service agreement. The company that emerges would be smaller in assets and lighter in capital intensity, with its most profitable divisions intact.

“The price is too high.” The price would be financed by bonds repaid from the network’s own revenue, sized by the same method Chapter 3 uses: what a well-run delivery network’s earnings could support as debt, at a rate reflecting bonds that carry no federal guarantee. That method produces a range of roughly 55 to 85 billion dollars, not a number asserted in advance. A large figure is not itself a reason for alarm. What matters is whether the asset’s revenue can carry the debt, and that’s a question this proposal’s method is built to test rather than assume.

“It gives government too much power.” The proposal would bind the government’s hands more tightly than Amazon’s are bound today. The network would operate at published, non-discriminatory rates, serving every shipper on equal terms, under a statutory universal-service obligation. The acquisition would replace discretionary private power over the conduit of commerce with regulated public obligation.

FedEx and UPS. Both carriers would have reasons to oppose. However, the status quo is not their friend either: both companies now compete against a vertically integrated rival that controls its own demand. Amazon has already passed both the Postal Service and the private carriers in parcel volume. A Delivery Network in public hands, billing published rates and favoring no marketplace, would be a more predictable competitor than the one they face now. The third-party access provisions would open volume to competitive dynamics rather than close it.

The pending antitrust litigation. Nothing in this document prejudges an outcome that belongs to the courts, in any of the three proceedings described in Chapter 3. The acquisition would be a purchase of assets that expressly declines to assume marketplace liabilities, leaving the litigation and any resulting exposure with the seller. Though the removal of the fulfillment lever might supply much of the structural relief the California and FTC actions seek, the case for Deliver America does not require Amazon to be found liable of anything in any of them.

9. Recommendations and Next Steps

The Legislative Vehicle

The goal is a single enabling statute: one bill that would charter the new department within the Postal Service, authorize the American Commerce Bonds, and convert the workforce as a class on the TSA model. That bill should not be the second legislative act, not the first.

The first should be a study act. Congress could direct an independent examination of the acquisition's premises before anyone is asked to vote on the acquisition itself: a commissioned valuation of the Distribution Network on the going-concern basis Chapter 3 sets out, an actuarial review of the workforce conversion Chapter 4 describes, and a financing analysis testing the repayment model in Appendix C against scenarios chosen by impartial analysts. The Government Accountability Office and the Congressional Budget Office exist for exactly this work, and a study act would be a modest ask for members not yet prepared to endorse the transaction.

A study act would also discipline this proposal's own claims. Every figure in this document is offered with its basis shown so that an independent review could confirm, correct, or reject it. We invite review from all interested parties.

The Window to 2028

The remainder of 2026 belongs to public launch and discussion across this document, DeliverAmerica.org, and coalitions we build.

In 2027, the work should turn legislative and analytical. The De Coster and California trials are scheduled for that year, and whatever they decide, the country's attention will be on the structure of Amazon's power over commerce while they run.

By 2028, the proposal should stand as a developed framework with an independent study underway or complete, available to any candidate or platform committee looking for a serious answer on infrastructure, rural service, and automation.

Once More, With Feeling

To the policymakers, staff, analysts, and organizers who have read this far, thank you. We are looking forward to your feedback. We have 3 small asks:

Engage the framework. Test the valuation, the workforce mechanics, the repayment arithmetic. What did we fail to consider?

Commission an independent study. A member of Congress could introduce the study act this year. A committee could request the GAO work without one. A foundation or university could fund an independent version tomorrow.

Begin the negotiation conversation. For people positioned to open a channel to Amazon decision makers, let’s talk.

EX MANIBUS POPULI

🤜📦🤛🏾

DELIVER AMERICA

Appendix A: Financial Modeling and Assumptions

This appendix states what is publicly known about the finances of Amazon’s delivery network, what cannot be known from any public source, and the single measurement that would resolve the difference. It makes no forecast. The network’s future rates would be set by public officials and its volumes committed by contract, so projecting them would amount to predicting decisions this proposal itself would place before Congress. What an honest appendix can do instead is map the territory, known and unknown.

A.1 The delivery layer (known)

In 2025, Amazon’s delivery operation carried about 6.9 billion packages in the United States. That figure comes from the U.S. Parcel Shipping Index published by Pitney Bowes, a shipping technology company whose annual count of American package volume is the standard reference for the industry. It made Amazon the largest package carrier in the country, ahead of the Postal Service at about 6.2 billion, ahead of United Parcel Service, ahead of FedEx.

What is that carriage worth? Measured at the per-package rates the index uses to compare carriers, moving those 6.9 billion packages represents roughly 33.4 billion dollars a year of delivery work, or about 4.84 dollars per package. That per-package figure is worth pausing on, because it is far below what the famous carriers charge: the index puts United Parcel Service near 15.86 dollars per package and FedEx near 17.04. Part of the difference is honest economics, since Amazon’s packages are small and light, averaging 1.8 pounds against more than 8 for the big carriers. Part of it is that Amazon has never had to charge anyone a real price for this work, for reasons the next section makes clear.

These are the best numbers in this appendix. The volume is a direct industry count. The dollar value is an estimate built on that count, and this paper labels it as such wherever it appears.

A.2 The fulfillment layer (unknown)

Delivery is only half of what would be acquired. The other half is fulfillment: the warehouses where goods are stored briefly, picked from shelves, packed into boxes, and handed to the delivery operation. Here the public record goes dark, and it is important to understand exactly why.

Every public company in America must file an annual financial report, called a 10-K, with federal regulators. The law requires the company to report results by its major divisions, which accountants call segments, drawn along the lines management actually uses to run the business. Amazon’s management runs 3 divisions: North America, International, and its cloud computing business. A delivery and fulfillment network is not one of them. It is treated as an internal cost of selling things, the way a supermarket treats its stockroom, and so it appears nowhere as a business of its own. From what the authors can determine Amazon is honestly following reporting rules written for a simpler era.

Three important numbers are missing:

First, the revenue is unfindable. The fees outside sellers pay Amazon for fulfillment are folded into a line called third-party seller services, 172.2 billion dollars worldwide, which also contains sales commissions and other charges, with no breakdown. The money Amazon earns delivering packages for anyone else entirely is folded into a line called Other, 5.9 billion dollars worldwide, which also contains health clinics, video licensing, and a credit card.

Second, the volume is unfindable. Amazon discloses what share of items sold come from outside sellers, but not the total number of items, and items are not packages: several items often travel in 1 box, and many of the boxes travel on other carriers.

Third, the costs are unfindable. Amazon reports 109.1 billion dollars of worldwide fulfillment expense and 102.7 billion dollars of worldwide shipping cost, but those totals also contain its retail stores, its customer service centers, its payment processing, and the money it pays the Postal Service, United Parcel Service, and FedEx to carry its packages. The cost of the network alone cannot be separated out.

A.3 Why Amazon’s own price lists do not fill the gap

Amazon publishes prices for its fulfillment services. Why not simply multiply the published prices by the volume?

Amazon does publish price lists, called rate cards, for two services of interest to us. One, Fulfillment by Amazon, is for sellers whose orders come through Amazon’s own marketplace. The other, Multi-Channel Fulfillment, is the identical warehouse work for orders that come from anywhere else. Comparing them is revealing. For the same size box at the same weight, the same pick from the same shelf packed by the same worker, the second service costs roughly double the first.

The reason is that neither price is really a price for fulfillment. The cheaper rate is subsidized by sales commission and advertising money on marketplace orders. The higher rate is defensive. Since 2026 the marketplace rates have even varied with the retail price of the item inside the box, something possible only because Amazon owns the marketplace and knows what every item sold for.

A.4 What the two networks look like combined

This proposal would join Amazon’s network with the Postal Service, so it is fair to ask what the combined operation looks like on paper today.

The two networks’ fortunes are not independent. For years, Amazon was the Postal Service’s largest package customer. Then it built its own delivery operation and, year by year, moved its own packages off the postal network and onto its own trucks. Amazon’s volume growth and the Postal Service’s volume decline are, to a substantial degree, the same packages changing hands. When Amazon’s count rose to 6.9 billion and the Postal Service’s fell to 6.2 billion, no new packages were created; carriage shifted from a public network to a private one. The Postal Service, losing volume, raised its average package price 11.2 percent in 2025, which is reasonable behavior for a network spreading fixed costs over fewer packages.

Combine the USPS and Amazon operations on paper, at their actual reported results, and the merged entity’s recent history would show a business whose two halves have mostly been trading market share with each other. It is a strong practical argument for the combination: the duplication of routes, facilities, and fixed costs that drives both sides’ economics.

A.5 The break-even model

The network’s true finances are hidden, but we can attempt to sketch the model a lender would: fix the terms Congress would control, and then calculate what the hidden numbers would have to be for the purchase to carry itself. The result is a set of conditions, each of which the diligence described in Appendix C would either confirm or refute.

The fixed terms: Bonds repaid over 30 years. Interest at 6 percent, a cautious planning rate for debt that carries no federal guarantee. A lender’s standard cushion: the network’s operating profit must exceed the annual debt payment by a quarter, a test the bond market calls 1.25 times coverage. And 3 candidate prices spanning the range this paper adopts: 55, 70, and 85 billion dollars.

From those terms, ordinary mortgage arithmetic produces the profit each price demands.

Purchase priceAnnual debt paymentOperating profit required
55 billion dollars4.0 billion5.0 billion a year
70 billion dollars5.1 billion6.4 billion a year
85 billion dollars6.2 billion7.7 billion a year

Profit is revenue multiplied by operating margin, the cents of profit earned on each dollar of revenue. Neither the network’s revenue nor its margin is publicly known. But for any pair of them, the table above says immediately whether the purchase works. The next table turns that around and asks: at each realistic margin, how much total revenue would the network need?

At 6 cents per dollarAt 8 centsAt 10 cents
To support 55 billion83 billion revenue62 billion50 billion
To support 70 billion106 billion79 billion64 billion
To support 85 billion129 billion96 billion77 billion

For calibration: United Parcel Service earns about 9 cents on the dollar, FedEx about 6, Amazon’s own North America division about 7, and the best-run freight carrier in the country 28. A margin of 6 to 10 cents is the realistic band for a well-run network; anything above that band should not be assumed and anything below it is the Postal Service’s current condition.

One known number anchors the revenue side. The delivery work alone, section A.1’s count, is worth roughly 33.4 billion dollars a year. So the tables above can be restated as a single question about the half of the business the public cannot see: how much revenue must fulfillment, the warehouse half, supply?

At 6 cents per dollarAt 8 centsAt 10 cents
For 55 billion50 billion from fulfillment29 billion17 billion
For 70 billion73 billion46 billion30 billion
For 85 billion95 billion63 billion44 billion

These are the numbers a member of Congress should hold while reading the diligence results. If the warehouses turn out to generate 30 to 45 billion dollars a year at a normal margin, the center of this paper’s range is affordable. If they generate 17 billion, only the low end is. If they generate less than that, the purchase as priced here does not close.

Three adjustments would move the thresholds:

A.6 Go and no-go: the conditions stated

The model reduces to a short list of conditions, each one a number the diligence would return, each one checkable against the tables above.

Diligence returnsReading
Network revenue of 79 billion dollars or more at a margin of 8 cents or betterThe 70 billion dollar center closes without a guarantee. Go.
Revenue of 62 to 79 billion at normal marginsThe 55 to 70 billion range closes; the price negotiates within it. Go, at the supportable price.
Revenue near 50 billion at strong margins, or 62 billion at thin onesOnly the low end closes unguaranteed. Go requires either a lower price, a Treasury guarantee, or longer terms. Conditional.
Margins below 6 cents at any revenue, or revenue below 50 billionThe purchase does not carry itself at any price in this paper’s range. No go, as priced, though the finding itself would be publicly valuable

These conditions are generous to no one - they assume no efficiency gains from combining the networks, no growth in package volume, no revenue from services the public network might add. Every one of those, if realized, moves the thresholds down. This model asks only whether the network, exactly as it operates today, could pay for itself.

A.7 The measurements, specified

The diligence provisions of the proposed legislation, described in Appendix C, would direct the government to obtain that number and the records needed to trust it. Stated as a specification, the required entries are the ones any buyer of any business would demand:

None of these numbers is exotic. Each exists now, in Amazon’s internal accounts, at whatever precision management uses to run the operation.

Appendix B: Workforce Data and Labor Cost Analysis

This appendix contains the workforce data and labor-cost analysis behind Chapters 4 and 5: workforce composition, turnover and compensation baselines, the federalization cost model, and pension mechanics.

Email for Access to Full Documentation

Appendix C: How To Pay For It

Wherein we explain how the United States could pay for the Delivery Network without spending tax money: what the law provides, what the law currently forbids, and what Congress must learn before setting a price. The tools described here are extant statutes, cited by section so that readers may convince themselves.

C.1 Known Unknowns

To understand whether an organization can carry debt, one must know what it earns and what it costs to run. Amazon’s delivery network is folded so deeply inside the rest of Amazon that those two numbers do not exist anywhere in public. Amazon reports its finances the way the law requires, in its annual SEC filing (form 10-K). That filing reports Amazon’s results by large divisions, and logistics is not one of them. That revenue is mixed into lines that also contain health clinics, video licensing, and a credit card. Its costs are mixed into worldwide totals that also contain retail stores and customer service centers.

So the question we must ask, whether the network’s own earnings could retire the purchase debt, depends on numbers only Amazon knows. No outside analyst can separate them. Congress, however, could compel their release.

C.2 Existing Machinery

One surprising finding of our research is that Congress has already built a financial vehicle this acquisition could use. It was built in 2006, and it has been sitting in the law, almost unused, ever since.

Some background:

The Postal Service sells 2 kinds of products under the law. Products where it has a monopoly, like ordinary letter mail, are market-dominant, and their prices are capped so a monopoly cannot gouge the public. Products where it competes with private companies, like package delivery, are competitive. The law treats competitive products differently: prices are not capped, but they must never be priced below cost, so a government service cannot undercut private companies unfairly. Packages are competitive products.

In 2006, Congress created a separate fund just for competitive products. It is called the Postal Service Competitive Products Fund, and it lives at section 2011 of title 39 of the United States Code. The fund has 4 properties that matter here.

Critics reasonably ask: would Amazon, as the network’s largest customer, simply dictate terms to its new public owner? The law says it cannot. Section 3633 requires that every competitive product cover its own attributable costs, meaning the costs that exist because that product exists. A federal regulator, the Postal Regulatory Commission, reviews contracts for compliance. Amazon’s negotiators could push for a low rate, but they could not push below cost, because a statute declares no contract can override it.

C.3 The Blocker

One provision of current law would need to be addressed in enabling legislation.

Section 2005 of title 39 caps the Postal Service’s total borrowing at 15 billion dollars. That ceiling was set in 1992 and has never been raised, not even for inflation. The same section limits how fast borrowing can grow: no more than 3 billion dollars of net new debt in any single year. Furthermore, the law explicitly adds borrowing under the Competitive Products Fund to the same 2 limits, so there is no side door.

An acquisition of this size cannot fit under a 15 billion dollar ceiling, and it cannot be drawn down at 3 billion dollars a year within any reasonable timeframe. The legislation this paper proposes would therefore amend both limits, expressly, as a named section of the bill. Congress wrote these caps and Congress can raise them.

C.4 The shape of the debt

Although the right price is not yet knowable, the cost of carrying any given price is ordinary arithmetic, the same kind used for a home mortgage. A borrower who takes a fixed amount at a fixed interest rate over a fixed number of years owes a fixed payment. The table below shows the annual payment for a range of possible prices, repaid over 30 years.

Financed price4% interest5% interest6% interest
55 billion dollars3.2 billion/yr3.6 billion/yr4.0 billion/yr
70 billion dollars4.0 billion/yr4.6 billion/yr5.1 billion/yr
85 billion dollars4.9 billion/yr5.5 billion/yr6.2 billion/yr

Which interest rate would apply? The law shapes the answer. Bonds issued under the Competitive Products Fund are, by default, not guaranteed by the United States, and the statute requires that this be printed on the face of each bond. Investors charge more interest for unguaranteed debt, and these bonds would also be taxable at the state and local level, which raises the rate further. This proposal therefore plans at 6 percent, the cautious end of the range. The law does contain a switch: under section 2006(c), the Treasury Secretary may, at the Postal Service’s request and in the Secretary’s own discretion, pledge the full faith and credit of the United States, which would lower the rate substantially. That decision would belong to the Treasury; we do not assume it.

For scale: Amazon’s delivery operation moved about 6.9 billion packages in 2025, more than any other carrier in the country. The estimated value of carrying those packages, at the rates the parcel industry uses to measure such things, is roughly 33.4 billion dollars a year. A debt payment of 4 to 6 billion dollars a year would need to come out of the profits of an operation of at least that size. Whether the profits are large enough is what we need to know.

C.5 The Boundary

Because the network’s true profit margin is unknown, this proposal does not name a purchase price. Instead, we name a method to produce a range.

The method is one any careful lender would use. Delivery is a thin-profit industry: the large American carriers earn between 6 and 10 cents of operating profit on each dollar of revenue, and the Postal Service currently earns less than zero. If the acquired network performs like a normal carrier, then its earnings can support a purchase price in the range of roughly 55 to 85 billion dollars, with the exact figure depending on its true revenue and margin. Appendix A presents this calculation in full. Prices far above that range would require the network to earn profits that no delivery company in America earns.

The final price would be an output of the process described below, and of negotiation with the seller, not a number asserted in advance.

C.6 Act I: Compelled Diligence

Every large acquisition begins the same way. The buyer announces its intent, and then, before money moves, examines the books. The examination is called due diligence.

The legislation this paper proposes would work identically. Its first title would direct the federal government to obtain the numbers that today exist only inside Amazon, using powers that, like the financing fund, already exist in law:

The Government Accountability Office, the auditing arm of Congress, already holds standing authority under section 2008 to audit postal accounts and report to Congress. The Postal Regulatory Commission already holds authority under section 2011(h) to establish, by formal rule, how the costs of the competitive business are identified and measured, and a 2022 law has directed it to refresh exactly those rules, a process that is live now. And the Commission already holds, under seal, the terms of the contract by which Amazon buys delivery from the Postal Service today, terms unknown to the public but fully known to the government.

The diligence title would direct these bodies, together with compelled disclosure from the seller, to produce: the network’s true operating profit; its revenue, separated from the rest of Amazon; its package volumes; its costs, isolated from businesses that would not be acquired; and its lease obligations. These are the entries any buyer of any company would demand.

The second title of the bill, raising the borrowing caps and activating the bond authority, would take effect upon the findings of the first. One bill, in sequence: examine, then execute.

C.7 The Decision

With diligence complete, Congress could set a supportable price and proceed, through negotiation with a willing seller, to a purchase.

It is worth stating what the alternative is. The federal government and 288 million American consumers are already in court with Amazon, in cases that will run for years, cost all parties enormous sums, and could end with a judge ordering the delivery network sold anyway, perhaps to private investment firms whose interest in universal service, worker welfare, and privacy is likewise not presumed. A negotiated purchase at a fair price resolves the same concerns without years of litigation expense, and with the seller compensated rather than punished.

The fund, the rate floor and audit power all exist. The constitutional mandate has existed since 1789. What would be novel is a decision by Congress to use them together.

Confidence scale:

1 = highest quality (primary source: statute, court opinion, GAO/OIG/PRC report, SEC filing, Amazon's own filing or release, Brookings).

2 = good secondary or a canonical landing page rather than the exact document (major news, industry tracker, agency home page).

3 = lowest quality (single local or regional report, advocacy source, aggregator, or a claim resting on secondary coverage of a primary that is not directly linked).

Where a link is a landing or search page rather than the exact document, it is marked "(landing)"; where a primary is paywalled or not directly located, an accessible secondary link is given and the rating reflects that.

Chapter 1, The Case for Public Acquisition

Amazon Supply Chain Services opened to all businesses (May 4, 2026); AWS-for-logistics framing; launch customers P&G, 3M, Lands' End, American Eagle; freight tier spanning ocean, air, ground, rail.

Amazon became the largest US parcel carrier in 2025; the crossover ran ahead of forecast.

USPS financial crisis figures (losses, cumulative, delivery points). Owned by Appendix C; see there.

Delivery Network volume growth of 688 percent, 2018 to 2024.

De Coster class of ~288 million consumers. Owned by Chapter 3; see there.

Chapter 2, Constitutional and Historical Foundation

The Postal Act of 1792 / Washington’s First Annual Address.

The Postal Reorganization Act of 1970.

Postal Clause and Necessary and Proper Clause (Article I, Section 8, Clauses 7 and 18).

Federal eminent domain for public use. Kohl v. United States, 91 U.S. 367 (1875).

Federal-Aid Highway Act of 1956 (41,000 miles; $25 billion authorization).

Tennessee Valley Authority Act (signed May 18, 1933); self-funding federal corporation.

Aviation and Transportation Security Act (November 19, 2001); TSA federalization; ~16,500 private to 40,000-plus federal screeners.

Rural Electrification Act (signed May 20, 1936); ~90 percent of farms without power.

Andrew Carnegie: sold Carnegie Steel to J.P. Morgan; $60 million; ~1,600 public libraries; 1892 Homestead Strike.

Chapter 3, Proposed Acquisitions (perimeter, valuation, litigation)

ASCS component map (AGL, SEND, AWD, MCD, PCP, FBA, MCF, Amazon Shipping) in Amazon's own names.

ASCS seller services page.

ASCS opened to all businesses; freight tier including rail.

Comparable valuations: UPS near $92 billion; FedEx near $78 billion (post FedEx Freight spin-off, June 1, 2026).

External shipping revenue $31.1 billion (2024), 12.2 percent YoY, 25.8 percent CAGR. Cross-reference Appendix A. Amazon 10-K underlies the fulfillment/shipping spend.

Facility and fleet inventory (600-plus US facilities, ~435 million sq ft, ~40,000 trucks, ~30,000 vans, ~110 aircraft).MWPVL International.

De Coster v. Amazon.com, Inc. (W.D. Wash.); ~288 million class; certified August 6, 2025; trial June 14, 2027.

California v. Amazon (AG Bonta, 2022; unredacted filing April 20, 2026; trial January 19, 2027).

FTC v. Amazon.com, Inc. Case No. 2:23-cv-01495, W.D. Wash., filed September 26, 2023. FTC joined by 17 states and Puerto Rico. Judge John Chun (same judge as De Coster). Bench trial rescheduled twice, now set for March 29, 2027 (from an original Feb. 9, 2027 date). Alleges both an anti-discounting scheme and a tying arrangement conditioning product placement on use of Amazon’s fulfillment network.

Chapter 4, Workforce Transformation

Amazon warehouse injuries above the warehousing-industry average (2017 to 2023), from Amazon's own OSHA data. Senate HELP Committee majority staff report, December 2024.

Time off Task (automated discipline from productivity logs).

State warehouse quota-disclosure laws (California, New York, Washington, Minnesota).

Delivery Service Partners; drivers are W-2 employees of partner firms.

Amazon Flex misclassification. Amazon Logistics, Inc. v. Virginia Employment Commission, 78 Va. App. 521 (2023), aff'd (Va. 2025, on procedural grounds); Wisconsin Supreme Court (via NELP, August 2025).

Coverage by employing entity; DSP health coverage at or above ACA standards.

TSA scale comparison; two-tier career/non-career postal employment. TSA cross-references Chapter 2. Two-tier: NALC and APWU national agreements.

OSHA enforcement over the Postal Service. Postal Employees Safety Enhancement Act of 1998 (Pub. L. 105-241).

Chapter 5, The Automation Dividend

More than 1 million robots (mid-2025); ~75 percent of delivery volume touches a robotic system.

Chapter 6, The Acquisition Structure

Six-day delivery requirement. 39 U.S.C. § 101(b), as amended by PSRA 2022.

USPS self-funding framework; Treasury borrowing; Negotiated Service Agreements; post-2022 non-postal-services authority. USPS OIG, "Business or Public Service?"

Universal-service obligation (every address; no small post office closed solely for a deficit). 39 U.S.C. § 101(b).

Bond-financing precedents (Liberty and Victory bonds; Highway Trust Fund; municipal revenue bonds).Illustrative.

Chapter 7, The Transition Plan

Synthesis chapter; most content cross-references Chapters 3, 4, 5, and 6.

Warehouse workforce employed principally by Amazon.com Services LLC.

APWU and NALC collective bargaining relationships.

TSA 2001 to 2002 class-conversion precedent. Cross-reference Chapter 2 (ATSA).

Chapter 8, Political Architecture

PSRA 2022 passed with substantial bipartisan majorities. PSRA 2022.

CSRS reallocation ~$75 billion, unaddressed for a decade. Cross-reference Chapter 8.

TVA self-funding without a congressional appropriation since 1959, 65-plus years. Cross-reference Chapter 2.

De Coster, California v. Amazon, and FTC v. Amazon.com; ~72 percent marketplace share (De Coster plaintiffs' estimate). Cross-reference Chapter 3.

Amazon Supply Chain Services opening access. Cross-reference Chapters 1 and 3.

FedEx and UPS as opponents. Characterization; no citation. If lobbying figures are added, source to https://www.opensecrets.org.

Chapter 9, Recommendations and Next Steps

GAO, CBO, and PRC as the bodies for a study, valuation, and diligence phase.

Amazon has reduced its postal business by 20 percent.

De Coster and California trials scheduled 2027; parcel-volume crossover 2025. Cross-reference Chapters 3 and 1.

Appendix A, Financial Modeling and Assumptions

Facility inventory (MWPVL). Cross-reference Chapter 3. Re-pull by type to confirm.

Appendix B, Workforce Data and Labor Cost Analysis

Benefits ~30 percent of compensation (private), ~38 percent (state and local). Bureau of Labor Statistics, Employer Costs for Employee Compensation.

Amazon among the larger employers of SNAP and Medicaid recipients; fewer than 10 percent of aid recipients work in the public sector. GAO-21-45 (and follow-up GAO-21-410T).

Turnover ~150 percent; ~$8 billion turnover cost. NYT 2021; Engadget October 2022.

Loaded compensation ~$29 an hour. WHYY, May 2025.

DSP driver population (~275,000 US, partner-firm W-2 employees). US-scoped estimate consistent with Amazon's global 390,000 DSP driving jobs.

Appendix C, How To Pay For It

USPS net loss $9.0 billion (FY2025), $9.5 billion (FY2024), $6.5 billion (FY2023); ~$109 billion cumulative since 2007; on the GAO High Risk List since 2009.

PAEA 2006 prefunding (~$5.5 billion a year); PSRA 2022 repeal and Medicare integration.

168 million delivery points, six days a week.

CSRS cost-allocation correction worth ~$75 billion in USPS's favor. PRC and USPS OIG findings.

Workers at large employers on federal assistance while employed. GAO-21-45.

AI Disclosure

In the creation of this document, I have used Claude (Sonnet 4.6, Sonnet 5.0, Opus 4.8 and Fable 5.0) as a research analyst and editor. While I am an informed voter and citizen, I have not previously made policy recommendations beyond the County level. I used Claude to:

The proposal's arguments and conclusions are my own. Errors are my own.

Throughout the process, Claude has shared insights from financial data, American history and current events. While not all insights were salient (or accurate), in some surprising instances they updated my understanding of the problem space and impacted the scope and content of this document. Claude is a promising intern with a bright future ahead of them!

Every citation has been (or is in the process of being) reviewed for accuracy by a human. Every released version was proofread and approved by me (and other humans).

Comments and corrections are welcomed at deliveramerica@deliveramerica.org

Katharine Lee Nelson

Portland, OR

July, 2026

a1789 postal act

↑