How The World Works | Fixed Costs

Before the first loaf is baked, the bakery already exists.

Rent has been signed.

The oven has been bought.

Licences have been paid.

The kitchen has been fitted.

Recipes have been tested.

The sign is on the door.

Then one loaf comes out.

If we allocate every setup dollar to that single loaf, it is the most expensive loaf imaginable.

Bake a thousand loaves and the oven does not need to be bought a thousand times.

The fixed setup cost is spread.

This is the world of fixed costs.

Quick Read

A fixed cost is a cost that does not change with the quantity of output over a specified short-run range.

Rent for a factory, a software development project, an insurance premium, a licence or a machine lease can behave as fixed costs over the relevant period.

Variable costs change with output: ingredients, packaging, some energy use, transaction fees or per-unit shipping.

OpenStax explains the central arithmetic:

average fixed cost = fixed cost ÷ quantity

As quantity rises, average fixed cost falls because the same fixed amount is spread over more units.

This creates powerful incentives to use installed capacity, grow volume and reach sufficient scale.

But fixed costs also create risk.

The cost remains even when demand falls.

The central question is:

What must be paid before or regardless of each additional unit—and how many useful units can that commitment support?

The One-Sentence Answer

Fixed costs work by creating an upfront or period cost that does not rise with each unit of output across the relevant range, so higher utilisation lowers fixed cost per unit while low utilisation makes the same infrastructure expensive.

Fixed Relative to What Time Horizon?

No cost is metaphysically fixed forever.

A building lease may be fixed this year and renegotiable next year.

A factory is fixed capacity this month and replaceable over ten years.

OpenStax distinguishes the short run, where at least some inputs are fixed, from the long run, where all inputs can eventually be varied.

So every fixed-cost statement needs a horizon.

Fixed for how long, and over what output range?

Fixed Cost Is Not Sunk Cost

These concepts are related and often confused.

A fixed cost does not vary with current output over the relevant period.

A sunk cost is a cost already incurred that cannot be recovered.

A monthly lease payment can be fixed but not yet sunk if the contract can still be terminated before future months.

Money already spent on failed research is sunk, even if the research cost was not fixed in an accounting sense.

The distinction matters for decisions.

Fixed cost affects future operating economics.

Sunk cost should usually not determine whether future spending is worthwhile.

The Fixed-Cost Chain

upfront commitment → installed capacity → fixed periodic cost → output volume → cost spreading → lower average fixed cost → pressure to utilise capacity → scale incentives and risk

The phrase “pressure to utilise capacity” is where economics becomes behaviour.

Once a costly asset exists, organisations want to keep it busy.

That can be efficient.

It can also distort decisions when utilisation becomes an objective even after the asset no longer fits the best strategy.

Average Fixed Cost

Suppose a machine costs $10,000 per month regardless of whether it makes one unit or ten thousand.

At 100 units, average fixed cost is $100 per unit.
At 1,000 units, it is $10.
At 10,000 units, it is $1.

The machine did not become cheaper.

The allocation of its cost changed because output changed.

This simple arithmetic is one of the foundations of economies of scale.

Variable Costs

Not everything spreads.

Each extra loaf needs flour.

Each parcel needs packaging.

Each transaction may carry a payment-processing fee.

These are variable or marginal costs.

Total cost is therefore shaped by both fixed and variable components.

A business with low fixed costs and high variable costs behaves differently from one with high fixed costs and low variable costs.

High Fixed Cost, Low Marginal Cost

Digital products make this cost shape unusually visible.

Building software can require years of engineering.

Serving one additional user may cost very little until infrastructure limits are reached.

A film is expensive to produce and cheap to show to another digital viewer.

A scientific dataset is expensive to collect and cheap to copy.

This cost shape encourages scale.

It also encourages very low marginal prices, subscriptions, bundling and freemium models.

The First-Copy Problem

Information goods often have a “first-copy” problem.

The first copy of a textbook is extremely expensive because it contains research, writing, editing, design and verification.

The millionth digital copy can be nearly free to reproduce.

This does not mean the content itself was free.

It means production cost and replication cost are radically different.

Public-goods-like knowledge often sits on top of this fixed-cost structure.

Fixed Costs Create Economies of Scale

If fixed cost is substantial, larger output can lower average total cost because each unit carries a smaller fraction of the fixed burden.

OpenStax explains this directly in its short-run cost discussion: average total cost begins high at low output because fixed cost dominates, then falls as fixed cost is spread.

This is not the only source of scale economies, but it is one of the cleanest.

See How The World Works | Scale.

Step-Fixed Costs

Fixed cost often stays flat only until a capacity threshold.

One supervisor can manage twenty workers.

At worker twenty-one, another supervisor may be required.

One server cluster handles one million requests.

At the next load band, another cluster is added.

One classroom serves thirty students.

At thirty-one, the school may need another class and teacher.

These are step-fixed costs.

Cost stays flat within a range, then jumps.

The Lumpy Investment Problem

Some capacity cannot be bought continuously.

You cannot buy 0.17 of an airport runway.

You cannot install half a minimum-size turbine.

You cannot hire 0.03 of a specialist if the job requires a full person at certain hours.

Indivisibility creates lumpy fixed commitments.

This is why capacity often arrives in steps and why utilisation can swing after expansions.

Utilisation Changes Unit Economics

A hotel pays for the building whether ten rooms or one hundred rooms are occupied.

An airline owns or leases the aircraft whether the flight is half full or nearly full.

A school pays for classrooms, administration and facilities even when enrolment falls.

Higher utilisation spreads fixed cost.

That creates powerful pricing incentives to fill spare capacity if the marginal cost of an extra user is low.

Why Airlines Discount the Last Seats

Once a flight is scheduled, many costs are committed.

The aircraft, crew, airport slot and basic flight cost exist whether one more seat is sold or not.

The marginal cost of one additional passenger can be far lower than average cost.

This helps explain why pricing can vary dramatically by timing, demand and seat availability.

The fixed-cost structure creates an incentive to monetise otherwise empty capacity.

Operating Leverage

High fixed costs create operating leverage.

Once revenue exceeds fixed cost, additional sales with healthy contribution margins can increase profit rapidly.

But when revenue falls, fixed costs remain.

Profit can collapse quickly.

High fixed-cost businesses therefore can have strong upside at high utilisation and severe downside at low utilisation.

The same cost structure that creates efficiency creates fragility.

The Break-Even Point

If each unit contributes a certain amount after variable cost, the firm must sell enough units for total contribution to cover fixed cost.

This creates a break-even volume.

High fixed cost pushes break-even volume upward.

High contribution margin pulls it downward.

The concept matters far beyond business.

A rail line needs enough ridership to justify infrastructure.

A specialist programme needs enough students.

A research instrument needs enough valuable experiments.

Fixed Costs Create Entry Barriers

If entering an industry requires enormous upfront capital, small entrants face a difficult problem.

They must pay the fixed cost before they have customers.

Incumbents already spread equivalent infrastructure across large output.

This can create barriers to entry in industries such as utilities, semiconductor fabrication, aviation, telecommunications or large-scale platforms.

Fixed cost therefore shapes competition.

Natural Monopoly Pressure

Some networks have such large fixed infrastructure costs that one provider can serve the market at lower average cost than several duplicate networks.

Water distribution and electricity grids have often displayed this structure.

Duplicating every pipe or wire can be wasteful.

But a single provider creates market power.

The fixed-cost problem therefore becomes a governance problem.

Efficiency of infrastructure and competition in service provision may need to be separated institutionally.

Fixed Costs and Density

Density determines how many users can share fixed infrastructure per unit area.

A kilometre of fibre serving ten households is expensive per household.

The same kilometre serving one thousand households spreads the trenching, equipment and maintenance base more widely.

Density can therefore turn infrastructure from infeasible to viable.

See How The World Works | Density.

Fixed Costs and Scale

Scale spreads fixed cost.

This is why high-fixed-cost industries often chase volume.

But scale can eventually require new fixed-cost layers.

A factory reaches capacity and another plant is needed.

A platform needs another data centre.

A school network needs regional management.

Fixed costs therefore create stepwise scale rather than one smooth decline forever.

See How The World Works | Scale.

Fixed Costs and Public Goods

Many public goods have high creation cost and low marginal cost.

A satellite weather system is expensive to build.

One more person reading the forecast costs little.

A scientific database is expensive to create and cheap to copy.

This cost shape reinforces the challenge of charging individual beneficiaries and the attractiveness of broad funding.

See How The World Works | Public Goods.

Fixed Costs and Opportunity Cost

Once capital is committed to one fixed asset, it cannot simultaneously fund another.

Building a stadium means not building something else with the same land and capital.

Opening a second branch means staff and money are unavailable elsewhere.

High fixed costs therefore deserve careful opportunity-cost analysis before commitment.

See How The World Works | Opportunity Cost.

Fixed Costs and Irreversibility

A fixed investment can be reversible or irreversible.

A standard machine may have a resale market.

A highly specialised factory may not.

The more irreversible the fixed commitment, the more valuable waiting and staged investment can become under uncertainty.

See How The World Works | Irreversibility.

Fixed Costs and Path Dependence

A large fixed investment creates reasons to remain on the chosen path.

Infrastructure is built.

Workers are trained.

Suppliers adapt.

Processes are optimised around the asset.

The original fixed cost becomes part of a larger complementary ecosystem.

See How The World Works | Path Dependence.

Fixed Costs and Substitution

High fixed infrastructure can make substitution slow.

A transport system built around roads cannot switch instantly to rail.

A company with a proprietary data centre cannot instantly become cloud-native without migration cost.

Fixed assets create installed-base inertia.

See How The World Works | Substitution.

Fixed Costs and Complementarity

An expensive asset may be worthless without complements.

A train line needs stations, signalling and feeder transport.

A hospital scanner needs trained staff and maintenance.

A school device programme needs connectivity, software and teacher capability.

The relevant fixed cost is often the cost of the working bundle, not the headline asset.

See How The World Works | Complementarity.

The Empty Building Problem

Infrastructure can exist without useful demand.

A school is built where population declines.

A shopping centre opens where footfall never arrives.

A factory is built for a product whose market disappears.

The fixed cost is now stranded.

Good planning therefore needs demand uncertainty and flexibility, not only engineering feasibility.

Stranded Assets

A stranded asset is an asset whose economic value falls unexpectedly because demand, technology, regulation or environmental constraints change.

High fixed-cost systems are exposed because capital is already embodied in infrastructure.

The risk becomes larger when the asset is highly specialised and difficult to repurpose.

Reversibility and modularity therefore reduce fixed-cost risk.

Why Leasing Exists

Leasing can convert part of a large fixed capital commitment into a more flexible periodic cost.

Cloud computing can do something similar for computing infrastructure.

Instead of buying a data centre, a firm rents capacity.

This can lower entry barriers and improve reversibility.

The provider still carries fixed infrastructure somewhere.

The fixed cost has been pooled and redistributed across many customers.

Infrastructure as Shared Fixed Cost

Roads, ports, grids, broadband and standards lower the fixed cost that each private actor would otherwise face.

A firm does not need to build its own port before exporting.

A household does not build its own power station.

Shared infrastructure socialises certain fixed costs and creates a platform on which many private activities can operate.

This is one deep economic role of civilisation.

Standards Lower Fixed Integration Costs

Without standards, every new connection can require custom engineering.

With a standard interface, the integration cost can be paid once and reused.

A plug standard, container size, data format or API contract therefore lowers repeated setup cost.

Standards convert bespoke fixed effort into reusable infrastructure.

See How Standards Work.

Automation Often Raises Fixed Cost and Lowers Variable Cost

A manual process needs more labour per unit.

An automated system requires machinery, software and engineering upfront.

After installation, labour per unit can fall.

Automation therefore often swaps variable cost for fixed cost.

This makes automation more attractive at high volume than low volume.

Why Small Firms Use General-Purpose Tools

A small firm cannot justify a custom machine for every process.

It uses flexible general-purpose tools with higher per-unit labour cost.

A large firm can justify dedicated equipment because the fixed cost is spread across more output.

This is why scale changes technology choice.

Fixed Costs and Nonlinearity

Average cost does not fall forever.

Fixed-cost spreading helps at first.

Then variable constraints, congestion or new capacity steps appear.

At a threshold, the organisation must buy another machine, building or team.

Cost jumps.

The unit-cost curve is therefore often nonlinear and piecewise.

See How The World Works | Nonlinearity.

Fixed Costs and Risk

A high-fixed-cost system needs demand.

If demand is uncertain, the commitment becomes risky.

One way to reduce risk is staged investment.

Build a smaller first plant.

Lease before buying.

Use modular data centres.

Pilot before national rollout.

Flexibility has option value when future utilisation is uncertain.

Fixed Costs in Education

Education has several fixed-cost layers.

A school building exists before the first lesson of the day.

A curriculum takes years to develop and can then be used across many classrooms.

A teacher spends time preparing a strong explanation that can serve several students.

A digital lesson can be expensive to research and produce but cheap for the next learner to access.

These layers explain why education can gain from shared infrastructure while still needing low-density human attention at the learner interface.

Lesson Preparation as Fixed Cost

A teacher may spend two hours preparing a lesson.

If the lesson serves one student, preparation cost per student is two hours.

If the same preparation can serve three students without losing fit, average preparation cost falls.

If it serves thirty students but individual feedback collapses, another cost appears.

Fixed-cost spreading therefore has to be balanced against density and quality.

Curriculum as Shared Infrastructure

A strong curriculum is expensive to design.

Once built, thousands of teachers can use it.

This is scale economy.

But curriculum cannot replace local diagnosis.

The fixed knowledge infrastructure should lower repeated design cost while leaving enough flexibility to respond to actual learners.

Fixed Costs in Science

A telescope, particle accelerator or genome-sequencing facility can be extraordinarily expensive.

Once built, many experiments can use the infrastructure.

Shared facilities spread fixed capital across researchers and institutions.

This is why scientific infrastructure is often nationally or internationally funded.

Fixed Costs in AI

AI systems can contain enormous fixed or quasi-fixed costs in model development, training infrastructure, evaluation, safety work and data preparation.

Inference then adds variable compute cost per request.

The economic structure therefore depends on both layers.

Large user bases spread development cost but increase total inference load, support, moderation and reliability requirements.

Scale changes the fixed-cost architecture again.

The “Use It Because We Bought It” Trap

A school buys expensive software.

A hospital buys a machine.

A company builds a headquarters.

Now people feel pressure to use the asset because it was expensive.

This can be rational if additional use has low marginal cost and real value.

It becomes sunk-cost fallacy when the asset no longer serves the best future choice.

Fixed-cost utilisation must not override current opportunity cost.

Make-or-Buy Decisions

Should a firm build a capability internally or buy it from a supplier?

Internal production may require fixed setup cost but offer lower variable cost at scale.

Outsourcing can avoid fixed commitment but charge a higher per-unit price.

Volume, uncertainty, quality, control and strategic importance determine which route wins.

This is fixed-cost reasoning inside organisational boundaries.

Fixed Cost Can Create Predatory-Looking Prices Without Predation

When marginal cost is low, a firm can rationally price additional units far below average historical cost while still improving contribution.

This can look surprising.

A nearly empty hotel room may be discounted heavily tonight because tomorrow the room-night disappears forever.

The relevant short-run comparison is incremental revenue versus incremental cost, subject to long-run strategy and price effects.

Average cost and marginal decision are different levels of analysis.

The Time-Sensitive Fixed Asset

An airline seat expires when the plane departs.

A hotel room-night expires at midnight.

A theatre seat expires when the performance begins.

These assets have fixed capacity and perishable inventory.

Revenue management exists because the system must allocate a fixed-capacity opportunity before time removes it.

Fixed Costs and Second-Order Effects

Large fixed investments change behaviour around them.

A highway encourages development.

A university campus attracts businesses.

A data centre attracts network infrastructure.

The fixed asset becomes a focal point for complementary investment.

That can improve utilisation and deepen path dependence.

See How The World Works | Second-Order Effects.

The Fixed-Cost Audit

Define the time horizon. Over what period is the cost fixed?
Define the output range. Within what capacity band does the cost stay unchanged?
Separate fixed from variable cost. Which costs rise with each unit?
Separate fixed from sunk cost. Which commitments can still be avoided or recovered?
Calculate average fixed cost. How does utilisation change cost per unit?
Find step costs. At what threshold does another fixed layer appear?
Find indivisibilities. Which capacities must be bought in lumps?
Calculate break-even volume. How much output is needed to cover fixed cost?
Check demand uncertainty. What happens if utilisation is low?
Check operating leverage. How sensitive is profit or viability to volume?
Check reversibility. Can the asset be sold, repurposed or scaled down?
Check complements. What else must be purchased for the fixed asset to work?
Check density. Are enough users located close enough to share the infrastructure?
Check scale. Does larger volume actually reduce average cost?
Check entry barriers. Does the fixed cost protect incumbents?
Check opportunity cost. What else could the capital fund?
Check utilisation pressure. Are people using the asset because it is valuable or merely because it exists?
Check second-order effects. What complementary development will the asset attract?

When the Fixed-Cost Lens Fails

The lens fails when “fixed” is treated as permanent.

Long-run decisions can change buildings, staff, contracts and machines.

It fails when fixed cost is confused with sunk cost.

It fails when managers chase volume without checking whether variable costs, congestion or demand quality make extra output worthwhile.

And it fails when low average fixed cost is treated as proof that the whole system is efficient.

A fully utilised bad asset is still a bad asset.

The Deep Decision Rule

Before commitment, ask whether the fixed asset deserves to exist.

After commitment, ask whether the next unit creates value above its relevant marginal cost.

Do not use the first question when answering the second.

Do not use sunk expenditure to justify future waste.

Fixed-cost intelligence is partly the discipline of knowing which cost belongs to which decision.

How Fixed Costs Connect to the Rest of the World

Scale: larger output can spread fixed cost and lower average cost.
Density: concentrated users make shared infrastructure easier to support.
Public goods: many shared services combine high creation cost with low marginal use cost.
Opportunity cost: fixed capital displaces alternative investments.
Irreversibility: specialised assets can be costly to unwind.
Path dependence: fixed infrastructure attracts complements and reinforces routes.
Substitution: installed assets can slow migration to alternatives.
Complementarity: a headline asset often needs an entire fixed-cost bundle.
Nonlinearity: step-fixed costs create jumps in the cost curve.
Capacity: fixed assets define short-run operating envelopes.
Competition: high fixed costs can create barriers to entry and concentration.
Second-order effects: infrastructure changes behaviour and investment around itself.
Risk: high fixed cost magnifies the consequence of demand volatility.

Questions a Reader Can Now Ask

Fixed over what time horizon?
Fixed over what output range?
Which costs actually vary with output?
Which costs are sunk and which are avoidable?
What happens to average fixed cost as utilisation rises?
Where is the next step-fixed threshold?
What break-even volume is required?
How uncertain is demand?
Can the asset be leased or shared instead of owned?
Can it be repurposed?
Does density support enough users?
Does scale lower unit cost enough to justify expansion?
What complements are missing?
Does the asset create an entry barrier?
Are we using it because it is valuable or because we already bought it?

Frequently Asked Questions

What is a fixed cost?

A fixed cost is a cost that does not change with output over a specified short-run range, such as certain rents, leases or installed infrastructure costs.

Is a fixed cost always sunk?

No. Fixed describes how cost behaves with output. Sunk describes whether a past cost can be recovered. A future lease payment can be fixed but avoidable; a failed research expense can be sunk without being a fixed operating cost.

Why does average fixed cost fall?

Because the same fixed amount is divided across more units of output.

Why are high fixed costs risky?

Because the costs remain when demand falls, creating operating leverage and a higher break-even requirement.

What is the most important practical rule?

Separate the decision to build the capacity from the decision to use already-built capacity. Before commitment, compare the full fixed investment with alternatives; after commitment, make incremental decisions using relevant future costs rather than sunk spending.

Research Basis and Further Reading

OpenStax, “Costs in the Short Run”, on fixed cost, variable cost and declining average fixed cost as output expands.
OpenStax, “Costs in the Long Run”, on scale, changing fixed-capital choices and economies or diseconomies of scale.
How The World Works | Scale, for the wider architecture created when fixed costs are spread across larger systems.

The Larger Idea

The first unit is rarely only the first unit.

It carries the oven.

The railway.

The classroom.

The server.

The software build.

The telescope.

The rules and training that make the whole apparatus usable.

Then the second unit arrives and the infrastructure does not need to be reinvented.

That is where scale begins to become cheap.

But the same infrastructure also becomes a commitment.

It wants users.

It wants volume.

It wants the future to resemble the assumptions made when it was built.

Sometimes it does.

Sometimes the world changes.

Fixed costs are the price of building a capability before you know exactly how much of the future will use it.

How The World Works | Fixed Costs — Why the First Unit Can Cost More Than the Thousandth