How Manufacturing Companies Make Money: The Complete Guide

Manufacturing companies make money by selling physical units for more than their full cost, which includes materials, the conversion that turns those materials into a product, and the plant, tooling and engineers paid for before anything is sold. Because so much of that cost is fixed, profit depends less on the margin printed on each unit than on how full the factory runs, how many units come out right the first time, and whether the balance sheet can carry the plant through a downturn. In many industries, the largest profits arrive after the sale, from parts, service and financing.

The third quarter of 2026 showed both sides of that business in two companies. Boeing delivered 161 commercial aircraft, its most in a quarter since 2018, and reported a record backlog of about $745 billion. Its Commercial Airplanes unit still lost about $697 million on about $10.72 billion of revenue (reported). In the same quarter, GE Aerospace reported orders of $19.0 billion, up 49%, from a backlog dominated by service contracts on engines already flying. One company was still paying for a factory it could not run at rate. The other was being paid for engines it had sold years earlier.

This guide explains where a manufacturer's money comes from, where it goes, and why some factories compound value for decades while others destroy it.

How do manufacturing companies make money?

A manufacturer sells a unit and pays for it three times. It pays once for materials: steel, resins, chips, castings and purchased parts. For cars, appliances and machines these are usually the largest cost, often more than half of the cost of goods sold. It pays a second time for conversion: the labor, energy, machine time, scrap and rework that turn materials into a finished product. And it pays a third time, in advance and whether or not the unit sells, for capacity: the building, the tooling, the salaried staff, maintenance and the engineering that designed the product.

The third payment is what separates manufacturing from trading. A distributor that buys and resells goods has mostly variable costs; if sales fall, so do its costs. A manufacturer's capacity costs stay. That single fact explains most of what makes the industry hard, and most of the analysis in Chapter 3 of The Production Equation.

Two manufacturers buying the same steel at the same price separate in two places. The first is conversion: how many labor hours, kilowatt-hours and scrapped parts each needs to make the same unit. The second is utilization: how well each fills its fixed assets. The rest of the business model, from pricing power to the aftermarket, sits on top of those two.

Where does a manufacturer's margin go?

The easiest way to see the economics is to follow one product from the customer's price back to the factory. The numbers below are invented and round, for illustration only. A machine sells for $10,000 to an end customer through a dealer.

Illustrative only; invented numbers, not any company's.

Three lessons come out of a stack like this. First, a cost change that looks small at the shelf is large in the factory. A $200 rise in the price of steel inside the machine is 2% of the customer's price but a quarter of the manufacturer's profit. That is why tariffs and commodity spikes hit manufacturers so hard, and why the book's worked example on a compressor finds that a 25% tariff on one imported motor takes a fifth of the maker's profit.

Second, the factory's own improvements are worth more than they appear. Cutting conversion cost by 10% barely moves the shelf price but raises this maker's profit by more than 11%. Most of a product's cost is locked in when its design is frozen, which is why design for manufacture and value engineering matter so much (see Chapter 12 on a product worth making).

Third, the overhead line is a fiction of the planned volume. At lower volume the same fixed cost is spread over fewer units, and the $800 of profit disappears quickly.

Why fixed cost decides manufacturing profits

Operating profit equals volume times contribution (price minus variable cost) less fixed cost. Break-even volume is fixed cost divided by contribution per unit. When profit is thin relative to contribution, a small fall in volume erases it. This is operating leverage, and it is the single most important number to understand about any factory. Our explainer on operating leverage in manufacturing works through the arithmetic.

The research behind the book uses a stylized plant with $300 million of fixed cost and $2,000 of contribution per unit. At 200,000 units it earns $100 million. A 10% fall in volume cuts profit by 40%, and at 150,000 units, 60% utilization, it earns nothing. The same mechanism explains Ford's roughly $19.5 billion of special charges announced on 15 December 2025 and Stellantis's roughly €22.2 billion of charges announced on 6 February 2026: both were the price of removing fixed cost built for electric-vehicle volumes that did not arrive.

Leverage also varies by industry. Semiconductor fabs, whose cost is dominated by the depreciation of tools and cleanrooms, sit at one extreme. Contract electronics assemblers, with leased buildings and flexible labor, sit near the other.

The four factors that decide whether a manufacturer creates value

The Production Equation argues that a manufacturer creates durable value only when four things are true at once, and that the factors multiply rather than add. If any one is zero, the result is zero.

Around the four factors sit two forces that behave like weather: the demand and input cycle, and policy and geopolitics. Neither decides whether a manufacturer is sound. Both decide how long an unsound one survives.

The multiplication matters because shortfalls compound. In the book's opening chapter, a stylized plant planned to earn $79.4 million. A 20% demand shortfall cut that to $13.5 million. A ten-point fall in first-pass yield turned it into a $45.0 million loss. Both together produced an $80.3 million loss, larger than either alone. Most of the failures in the book come from treating one factor as proven because another was: a subsidy mistaken for a product worth making, a pre-order list mistaken for demand, automation mistaken for process capability, a tariff mistaken for a competitive cost.

Why is process capability the hardest factor to buy?

A product can be licensed, a plant bought and capital raised. Process capability lives in people, standard methods, tooling and data, and it takes years to build. It also decays quietly. A rate increase that outruns training, a supplier squeezed until it cuts inspection, a rework step nobody records: none shows in the quarter's numbers, and all show up later as a recall, a regulator's cap or a lost customer.

The Toyota Production System is the best-known way of building that capability, and NUMMI is its proof. The Fremont plant General Motors closed in 1982 reopened in 1984 as a joint venture with Toyota, with largely the same union workforce, and within a few years reached quality and productivity comparable with Toyota's plants in Japan (MIT, reported). The machines were not the difference. The system was.

Boeing shows the cost of losing it. On 5 January 2024 a door plug blew out of a ten-week-old 737 MAX 9 because four bolts had not been reinstalled at the factory. The FAA capped 737 output at 38 a month and did not allow 42 until October 2025. Our case study on what went wrong at Boeing traces how a company with a record backlog lost control of its own production rate.

Capability is also what the most quoted factory metric tries to measure. Overall equipment effectiveness multiplies availability, performance and quality; our guide on how to calculate OEE shows why the same line can honestly report 65% or 86% depending on definitions, and why the cheapest capacity a manufacturer can buy is usually the time its own lines already waste.

Where do manufacturers earn most of their profit?

For many capital-goods makers, the answer is after the factory gate. Engine makers sell new engines at a loss or a thin margin and earn their returns from decades of parts and overhauls. GE Aerospace delivered 1,802 LEAP engines in 2025, up 28%, and said it expected the original-equipment business to turn profitable only in 2026. Its backlog of more than $210 billion at mid-2026 included about $170 billion of commercial services (company).

Caterpillar shows the same pattern in machinery. Its services revenue was about $24 billion in 2025, against a 2030 goal of $30 billion (company), earned from machines sold years earlier through a dealer network that rivals cannot easily copy. The book's section on aftermarket services and the installed base treats the installed base as an asset in its own right: the larger and longer-lived the fleet, the more predictable the tail.

Pricing power follows the same logic. A manufacturer can pass on costs when customers cannot easily switch: a certified aircraft part that would have to be requalified, a brand, or a specification written around its product. It cannot in commodities, where the price is the market's and the only defense is to be the low-cost producer. Nucor built its business on the second path, with minimills and a pay system that flex with the steel cycle.

Mix matters as much as volume. Full-size pickups and large SUVs have long been reported to earn several times the profit of small cars for similar labor hours, which is why the Detroit Three's US plants concentrate on trucks.

Make, buy or rent the factory?

Every manufacturer decides what to make and what to buy, and the decision turns on capability, control and capital more than on piece price. Outsourcing moves cost and risk, not just work. Boeing sold its Wichita fuselage plant to Onex in 2005 for about $1.2 billion (reported); twenty years later it bought the same operation back as part of Spirit AeroSystems, at a value of about $8.3 billion including debt (reported), because the fuselage line had become the constraint on its output.

At the other end sits contract manufacturing, an industry whose product is other companies' production. Foxconn, the largest, earns operating margins of roughly 2 to 3% on very large revenue by pooling demand across customers to keep its plants full. Our comparison of contract manufacturing vs in-house production covers when each model wins.

How capital and the cycle make or break manufacturers

A manufacturer commits capital years before it earns a return. A new aircraft program, chip fab or steel mill takes years from decision to first unit and more years to reach full output, and early units cost far more than later ones. Under program accounting, Boeing still carried about $13.86 billion of 787 deferred production costs at the end of 2025, fourteen years after the first delivery (10-K).

The cycle makes that commitment dangerous. A small fall in consumer demand becomes a larger fall in orders for steel, chips and machine tools as each tier destocks at once. Capital decided at the top often arrives in the trough. General Electric bought Alstom's power business at the top of the gas-turbine cycle, closing on 2 November 2015, and three years later cut its quarterly dividend to one cent. Our case study on why GE broke up follows that story to its end in 2024, when GE Vernova was spun off and what remained became GE Aerospace.

Pensions are the hidden lever. When Bethlehem Steel failed, the Pension Benefit Guaranty Corporation took over its plan, covering about 95,000 people, at a recorded loss of about $3.6 billion. The book's section on capital cycles, conglomerates and break-ups treats a pension deficit as what it is in substance: debt.

How do subsidies and tariffs change manufacturing economics?

Public money and trade policy moved more money through US manufacturing in 2025–26 than in any recent period. The CHIPS Act funded new fabs, and on 22 August 2025 the US government took about 9.9% of Intel, 433.3 million shares for about $8.9 billion, largely by converting grants it had already promised. Yet Intel's Ohio site, where ground was broken in 2022, is not expected to operate before 2030–31. A subsidy lowers the cost of capital. It cannot supply the customer that fills a plant. Our case study on why Intel Foundry is losing money explains why.

Tariffs work the same way. They can make a domestic product competitive for a time, but the legal basis of broad US tariffs changed four times in 2025–26, including the Supreme Court's 20 February 2026 ruling that the emergency-powers statute does not authorize tariffs. A plant that pays only behind a tariff is exposed to whoever changes it next. Our guide to reshoring and total landed cost shows why the labor gap, not the tariff, is what reshoring has to beat, and why credible reshoring is usually automation plus a tariff rather than a tariff alone.

Why do manufacturing companies fail?

Manufacturers rarely fail because their machines are bad. They fail when one of the four factors goes to zero while management believes another factor has covered it. Bethlehem Steel failed on capital, carrying pensions for a workforce far larger than the one still working. Takata failed on process capability in the field. Lordstown Motors bought a car plant on the strength of pre-orders that were not binding. Wolfspeed built a fab ahead of a demand curve that slowed and filed for Chapter 11 on 30 June 2025. Our list of the biggest manufacturing failures sorts them by the factor that broke.

The book's research on manufacturing failures, bankruptcies and busts adds a practical point about who pays. In GM's and Chrysler's 2009 bankruptcies, shareholders were wiped out, bondholders took equity at a discount, dealers were terminated and the Treasury's auto program closed with a net loss of about $9.3 billion (reported).

How to read a manufacturing company

The habits below work for a plant, a division or a listed company.

Start with contribution, then fixed cost. Find the break-even utilization and compare it with the industry's last trough. If the plant loses money below 75% and troughs run at 60%, the balance sheet must carry it through.

Separate announced, committed, spent and operating. Pledged investment is not spent investment, announced jobs are not hired jobs, and groundbreaking is not production. Foxconn announced a $10 billion plant in Wisconsin with up to 13,000 jobs in 2017; its contract was renegotiated in April 2021 to about $672 million of investment and 1,454 jobs.

Read quality as a cost line. Warranty expense as a share of revenue over several years, recalls and regulator findings say more than certificates. Stellantis's February 2026 reset included a €4.1 billion change in warranty provisions that management tied to quality problems.

Follow the margin past the factory gate. Parts, service, financing and software often earn more over a product's life than the original sale.

Watch cash, inventory and deferred balances, not just margins. Under absorption accounting, a plant that builds more than it sells reports a better profit than one that matches output to demand. The cash is gone either way.

Ask who carries the rate risk. A supplier on fixed prices absorbs its customer's rate changes until it cannot, and then the risk moves back to the customer, as it did when Boeing bought Spirit.

Operating leverage, the cycle and the four factors apply in every row of the industry table above. The book's preview chapters apply them in turn to cars, aircraft, machinery, chips, chemicals, steel and consumer goods.

The takeaway

A manufacturer makes money when it sells a product customers want at a price above full cost, makes it right the first time, fills its fixed assets and finances itself for the bad years as well as the good ones. Those four conditions multiply, so excellence in three cannot rescue a zero in the fourth: Boeing had demand, capacity and capital in 2024 and was still held back by its process. The industry's profits flow to whoever holds the scarce part of the chain at the time: engines in service, a working fab, a certified process or steel behind a tariff wall. The Production Equation is not yet released, but the book page and chapter previews set out the full framework, starting with Chapter 1, "Why Making Things Is Hard to Make Pay".