Corporate investment is normally expected to follow demand.
Orders rise, factories become busier, capacity utilisation increases and companies eventually decide that they need another production line, warehouse, data centre or manufacturing facility.
But in several important industries, that sequence is beginning to change.
Businesses are increasingly having to make capacity decisions before demand fully materialises. The reason is not necessarily excessive optimism. It is that the time required to build new productive capacity can be much longer than the time customers are willing to wait for it.
Factories take years to design and construct. Specialist machinery can have lengthy delivery schedules. New employees need to be recruited and trained. Grid connections can take years to secure. Suppliers themselves may have limited production capacity.
Companies that wait until demand is clearly visible can therefore discover that it is already too late to respond.
This is creating a new strategic problem: the capacity constraint.
In sectors ranging from semiconductors and electrical equipment to batteries, data centres and advanced manufacturing, businesses increasingly need to decide how much capacity they will require several years into the future.
That means investing before the orders necessarily exist.
The Traditional Capacity Model Is Being Challenged
For most businesses, adding capacity too early is dangerous.
A company that builds a factory without sufficient demand can be left with expensive assets producing little revenue.
Fixed costs continue.
Depreciation continues.
Employees still need to be paid.
Debt may still need servicing.
For that reason, management teams have traditionally preferred to see convincing evidence of future demand before committing large amounts of capital.
The difficulty is that this approach works best when capacity can be added relatively quickly.
If a company can install another production line within several months, waiting for demand is reasonable.
If increasing capacity requires three, four or five years, the calculation changes.
By the time demand becomes obvious, the company may already face a multi-year shortage.
Recent analysis from McKinsey on expanding US manufacturing capacity illustrates the problem. The firm notes that major manufacturing investments can take years to translate into actual production. Taiwan Semiconductor Manufacturing Company's Arizona facility, for example, was announced in 2020 but did not begin high-volume production until late 2024.
The implication extends well beyond semiconductors.
Capacity decisions increasingly need to anticipate demand rather than simply react to it.
Long Lead Times Change Corporate Behaviour
A lead time is effectively the distance between a company's investment decision and the moment new capacity becomes usable.
The longer that distance becomes, the more forecasting risk management must accept.
Consider industrial equipment.
A manufacturer experiencing rapidly rising orders may want to expand production, but building another facility is only part of the challenge.
It may also need:
specialised machinery,
electrical equipment,
automation systems,
trained technicians,
grid connections,
permits, and
additional suppliers.
Each element has its own lead time.
McKinsey's work on technology-focused capital projects has found that substations and transformers required for major facilities can face procurement lead times of roughly 50 to 150 weeks.
That creates a strong incentive to place orders well before the completed facility needs the equipment.
Companies are effectively reserving future capacity within their own supply chains.
The Electricity Grid Shows the Problem Clearly
Few sectors demonstrate the capacity constraint more clearly than electricity infrastructure.
Demand for electricity is rising in many markets as data centres expand, industrial processes electrify and new generation capacity connects to grids.
But the physical equipment required to expand those grids cannot necessarily be produced quickly.
The International Energy Agency's analysis of transmission-grid supply chains found that procurement times for major grid components have increased substantially. Large power transformers can now require up to four years to secure, while some cables can take two to three years.
Prices and waiting times for transformers and cables have roughly doubled since 2021.
The strategic consequence is obvious.
Utilities and infrastructure developers cannot wait for an electricity shortage before ordering equipment.
They must anticipate where demand will appear years in advance.
The same constraint affects companies building data centres, factories and other electricity-intensive facilities.
A company may have sufficient land and financing for a new project but still be unable to operate it if the necessary power infrastructure is unavailable.
Capacity planning is therefore spreading beyond individual companies into entire industrial ecosystems.
Data Centres Are Creating a New Industrial Capacity Cycle
Artificial intelligence provides one of the most visible examples.
Most public attention has focused on demand for advanced chips.
But AI infrastructure requires far more than semiconductors.
Data centres need transformers, switchgear, cooling systems, backup power, construction materials and enormous amounts of electricity.
The challenge is that many suppliers serving this market historically operated in slower-moving industries.
Transformer manufacturers, cooling-equipment suppliers and other industrial businesses often designed production systems around comparatively predictable demand.
AI infrastructure is changing those expectations.
Orders can expand much faster than manufacturing capacity.
This encourages industrial companies to increase production before every future data-centre project has been finalised.
Waiting for complete certainty risks losing the market to competitors that expanded earlier.
Semiconductor Factories Require Decisions Years Ahead
Semiconductors provide an even more extreme example.
Advanced fabrication plants are among the most complicated manufacturing facilities in the world.
They require enormous capital commitments, highly specialised equipment and skilled labour.
A fabrication plant cannot simply be added when chip demand suddenly increases.
McKinsey estimates that around $1 trillion could be invested in semiconductor capacity through 2030.
Those investments must necessarily anticipate demand several years into the future.
This creates a recurring industry challenge.
If manufacturers invest too little, chip shortages can emerge.
If they invest too aggressively, the industry can end up with excess capacity when demand weakens.
The objective is therefore not perfect forecasting.
Perfect forecasting is impossible.
The challenge is determining which demand trends are structural enough to justify building capacity before utilisation is guaranteed.
Battery Manufacturing Faces the Same Timing Problem
Battery production illustrates another version of the constraint.
Electric vehicles, grid storage and other applications are expected to require substantially greater battery capacity over the coming decade.
But new supply chains take time to build.
According to McKinsey's 2026 analysis of the Western battery supply chain, building battery capacity can require investment of around $1 billion for every ten gigawatt-hours, while major projects can have construction lead times of four to five years.
That creates an unavoidable forecasting problem.
A battery plant expected to begin production near the end of the decade must be financed and built years earlier.
Companies cannot know precisely how EV adoption, battery chemistry or prices will develop by then.
But waiting for clarity would mean capacity arriving too late.
The result is a form of calculated pre-investment.
Supply Chains Are Encouraging Companies to Reserve Capacity
The capacity constraint is not limited to factories companies own themselves.
It also affects suppliers.
A business planning a major expansion may discover that the most important bottleneck is a component manufactured elsewhere.
This can change procurement behaviour.
Companies increasingly seek longer-term agreements with critical suppliers.
Some reserve production slots.
Others commit to minimum purchases.
In certain industries, customers may even provide financing or investment support to suppliers in order to secure future availability.
This represents a shift from traditional purchasing.
Historically, procurement teams often focused heavily on reducing unit costs.
In capacity-constrained markets, availability can become more valuable than obtaining the lowest possible price.
A component that costs 5% less has limited value if it arrives two years after the factory needs it.
Manufacturing Capacity Is Becoming Strategic
Capacity has traditionally been treated primarily as an operational issue.
Factory managers worried about utilisation.
Operations teams monitored throughput.
Finance departments evaluated capital expenditure.
Increasingly, however, capacity is becoming a strategic concern for senior management.
Deloitte's 2026 Manufacturing Industry Outlook points to continued investment in smart manufacturing as companies seek greater productivity, resilience and flexibility. Deloitte's survey of manufacturing executives found substantial planned spending on automation hardware, analytics, sensors and cloud systems.
Part of this investment is designed to create more output from existing facilities.
That matters because building another factory is not always the fastest solution to a capacity shortage.
Sometimes the cheapest capacity is the capacity already sitting inside the plant.
Companies Are Trying to Unlock Existing Capacity First
Before building a new facility, companies increasingly examine whether existing assets can produce more.
Automation can reduce bottlenecks.
Predictive maintenance can reduce downtime.
Software can improve production scheduling.
Better supply-chain coordination can prevent machines from sitting idle while waiting for components.
Workforce changes can increase operating hours.
This is important because new factories are expensive and slow.
McKinsey's research on manufacturing ramp-ups suggests that significant production increases can sometimes be achieved by making greater use of existing capacity before building entirely new facilities.
This creates a hierarchy of capacity decisions.
First, improve existing operations.
Then expand existing sites.
Only after those options become insufficient does entirely new construction necessarily become the best choice.
Building Early Carries Real Risks
Investing before demand arrives is not automatically a good strategy.
It can go badly wrong.
Forecast demand may never appear.
A competitor may introduce a better technology.
Prices may fall.
Customer preferences may change.
A new plant may become obsolete before it has generated an acceptable return.
The semiconductor industry has repeatedly demonstrated how quickly shortage can turn into excess supply.
Battery manufacturing carries similar risks if capacity expands faster than vehicle demand.
Data-centre investment could also become vulnerable if expected AI workloads, efficiency improvements or power constraints change the economics of construction.
The capacity constraint therefore forces companies to balance two opposing risks.
Underinvestment risk: the company lacks enough capacity when demand arrives.
Overinvestment risk: the company builds capacity that customers never require.
Neither is trivial.
The Cost of Being Late Is Increasing
Historically, overcapacity often appeared to be the more serious problem.
Empty factories destroy returns.
But in some modern markets, being late can also be extremely expensive.
If a supplier cannot fill customer orders, buyers may redesign products around competitors.
If a utility cannot deliver a grid connection, industrial projects may be built elsewhere.
If a data-centre equipment manufacturer cannot deliver cooling systems, developers may establish long-term relationships with another supplier.
Capacity shortages can therefore permanently alter market share.
The opportunity cost is not only lost sales today.
It can be the loss of a customer relationship lasting many years.
This makes strategic capacity buffers more defensible.
Utilisation Is No Longer the Only Metric
Businesses have traditionally valued high capacity utilisation.
A factory running close to full capacity appears efficient.
Fixed costs are spread across more output.
But extreme utilisation can reduce resilience.
A factory running at 99% capacity has almost no room to respond to an unexpected increase in orders.
A supplier outage can quickly create shortages.
Maintenance becomes harder to schedule.
The same applies to logistics networks, data centres and energy systems.
A small amount of spare capacity can therefore have strategic value.
Companies increasingly need to distinguish between idle capacity and option capacity.
Idle capacity produces no economic benefit.
Option capacity creates room to respond quickly when conditions change.
The two can look identical on a utilisation spreadsheet but have very different strategic value.
Labour Can Be the Constraint Even When Machines Are Available
Capacity does not consist only of physical equipment.
People matter.
A company can build a factory but still struggle to operate it if skilled workers are unavailable.
Technicians, engineers, electricians, welders and specialist operators may require years of experience.
This means workforce planning must begin before new facilities open.
McKinsey's work on new factory construction highlights "time to talent" as an important constraint that can delay a plant from reaching full operating capacity even after physical construction progresses.
This reinforces the broader lesson.
Capacity has to be considered as a system.
Land without electricity is not capacity.
Machines without workers are not capacity.
Factories without suppliers are not capacity.
Capital Allocation Becomes More Difficult
Investing before demand arrives also creates a finance problem.
Management must commit capital today against cash flows that may not arrive for several years.
That can reduce near-term free cash flow.
Depreciation eventually increases.
Financing costs may rise.
Investors may question spending on factories that are initially underutilised.
The business case therefore needs to distinguish between temporary excess capacity and permanently unnecessary capacity.
A facility operating at 60% utilisation immediately after opening may not necessarily represent a poor investment if demand is expected to fill it over the following years.
But that argument can also be used to justify weak projects.
Management teams therefore need credible demand assumptions, clear milestones and disciplined return thresholds.
Capital Projects Themselves Are Becoming a Bottleneck
The construction process introduces another layer of risk.
Large factories often run over budget and behind schedule.
McKinsey's analysis of major factory projects found that large capital projects have historically experienced significant schedule delays and cost overruns.
This makes early investment even more complicated.
A company anticipating demand in 2029 might begin construction assuming a three-year project.
If delivery slips by 18 months, the capacity can still arrive late despite years of advance planning.
Businesses are therefore paying greater attention not only to whether they should build, but also to how reliably they can deliver major projects.
Geographic Capacity Is Also Changing
Another reason companies are building ahead of demand is supply-chain resilience.
Businesses are reassessing where critical products are manufactured.
Some industries have highly concentrated supply chains.
A disruption in one country or supplier can therefore affect customers globally.
Companies are responding by considering additional manufacturing locations, regional production and supplier diversification.
This capacity may appear redundant when global trade operates smoothly.
But redundancy can function as insurance.
The economics resemble spare capacity within a factory.
Operating everything at maximum efficiency can reduce costs in stable conditions.
A more diversified network may cost more but perform better when disruption occurs.
The Grid Shows Why Timing Matters
Electricity infrastructure provides a useful illustration of why anticipatory investment can be necessary.
The IEA's Electricity 2026 report describes grid capacity as an increasingly important bottleneck connecting new generation, storage and electricity demand.
Traditional grid projects can require many years.
New high-voltage lines can take seven years or more.
If utilities wait for actual electricity demand before beginning expansion, infrastructure can arrive far too late.
The IEA also estimates that technologies such as dynamic line ratings, power-flow controls and reconductoring can unlock substantial capacity more quickly.
This demonstrates another important principle of capacity strategy.
Companies do not always need to build from scratch.
Sometimes technology can unlock capacity from assets already in place.
The Winners May Be Suppliers to the Capacity Build-Out
The capacity constraint creates opportunities far beyond the companies building factories.
It supports demand for:
industrial machinery,
automation systems,
transformers and switchgear,
cooling equipment,
construction services,
engineering firms,
factory software,
industrial sensors, and
maintenance services.
These businesses effectively sell the tools required to create additional capacity.
In certain cases, they may benefit regardless of which end-market competitor eventually wins.
A data-centre developer faces the risk of selecting the wrong location.
A transformer manufacturer may sell equipment to multiple developers.
A semiconductor company must predict which chips will be required.
A supplier of specialist fab equipment can potentially serve several manufacturers.
The capacity cycle can therefore create an ecosystem of secondary beneficiaries.
Scarcity Can Encourage More Scarcity
Capacity constraints can also become self-reinforcing.
When companies fear equipment will become unavailable, they order earlier.
Earlier ordering increases suppliers' backlogs.
Longer backlogs encourage other customers to place orders even sooner.
This can create apparent demand well ahead of actual consumption.
The risk is that some orders are precautionary rather than permanent.
If customers later cancel or postpone projects, suppliers that expanded aggressively may suddenly face excess capacity.
This is another reason management teams need to distinguish between genuine structural demand and temporary scarcity behaviour.
Companies Are Buying Time
Ultimately, capacity investment is increasingly about time.
Businesses are not simply purchasing factories or equipment.
They are purchasing the ability to respond when future demand appears.
A company that begins expanding three years early may initially look inefficient.
But if competitors begin only after the market becomes visibly constrained, that early investment can become extremely valuable.
This does not eliminate the danger of forecasting incorrectly.
It changes the balance between waiting and acting.
The New Capacity Question
The traditional corporate question was:
How much capacity do we need today?
Increasingly, companies need to ask something different:
How much capacity must we begin creating today so that it exists when we need it?
That distinction is becoming increasingly important in industries where the physical economy cannot expand at digital speed.
Factories take time.
Transformers take time.
Semiconductor fabs take time.
Workers take time to train.
Electrical grids take time to expand.
The result is a strategic paradox.
Companies may have to invest before demand arrives precisely because waiting for demand could mean missing it.
In capacity-constrained markets, therefore, the greatest risk is not always building too early.
Sometimes it is discovering, several years too late, that the capacity required to compete was never built at all.
References
International Energy Agency — Building the Future Transmission Grid
McKinsey & Company — Semiconductors: Etching the New Map of Strategic Supply
McKinsey & Company — Incentive Pricing for Batteries: Scaling the Western Supply Chain
McKinsey & Company — Accelerating the Delivery of Tech-Focused Capital Projects
McKinsey & Company — Smarter Growth, Lower Risk: Rethinking How New Factories Are Built
S&P Global — Behind the AI Boom: The Electronics Supply-Side Constraints