Scarce production slots, computing power, energy connections and specialist manufacturing are turning future access to capacity into something companies increasingly secure before they need it.
For decades, procurement strategy was built around a familiar assumption: if a company had enough money, it could usually buy the capacity it needed. Factories could add shifts, cloud providers could provision more servers, airlines could order aircraft, manufacturers could source extra components and energy users could request larger connections. Price mattered, but availability was often treated as a secondary variable.
That assumption is becoming less reliable. In a growing number of industries, the scarce asset is not simply the product. It is the right to receive the product, use the infrastructure or occupy the production slot at the moment it is needed. Semiconductor fabrication, advanced packaging, AI computing, grid connections, specialist pharmaceutical manufacturing, aircraft delivery positions and even some categories of industrial equipment increasingly involve long lead times and limited near-term supply.
As a result, companies are beginning to treat contracted capacity as a strategic asset. They are signing multi-year commitments, making prepayments, accepting take-or-pay structures and reserving supply before demand is fully visible. The objective is no longer only to secure a lower price. It is to secure the ability to operate, grow or launch when competitors may still be waiting for access.
This changes the economics of resilience. Contracted capacity can reduce execution risk and protect revenue opportunities, but it can also lock up cash, reduce flexibility and leave companies paying for capacity they no longer need. The strategic question is therefore not whether to reserve capacity, but where scarcity is durable enough to justify putting balance-sheet resources behind access.
Capacity scarcity is moving upstream
The scale of the constraint is visible in electricity infrastructure. The International Energy Agency’s Electricity 2026 grid analysis estimates that more than 2,500 GW of renewable, storage and large-load projects are stalled in grid connection queues worldwide. It says new grid infrastructure can take five to fifteen years to plan and build, while prices for key grid components have nearly doubled over the past five years.
For a power-intensive factory or data centre, that means a site with land but no credible electricity connection may not be commercially equivalent to a site with contracted power capacity. The connection itself can become part of the asset value. A company that has already secured a megawatt allocation, substation slot or transformer order may be able to execute years earlier than a rival that starts the process after demand appears.
The same issue appears in AI infrastructure. The IEA’s Energy and AI analysis notes that wait times for critical grid components such as transformers and cables have doubled in recent years, while some generation equipment carries multi-year lead times. When compute facilities are constrained by both chips and power, access to one without the other does not create usable capacity.
This is why contracted capacity increasingly spans several layers at once. A cloud provider may need accelerator chips, data-centre construction, electricity supply, network equipment and cooling infrastructure to arrive on a coordinated schedule. An industrial manufacturer may need machine tools, qualified suppliers, logistics space and power upgrades. A pharmaceutical company may need validated production lines and regulatory-approved manufacturing capacity. Scarcity in any one layer can delay the whole system.
Semiconductors show how capacity becomes strategic
Advanced semiconductors are a clear example because fabrication capacity is expensive, technologically specialised and slow to build. In its 2025 annual report, TSMC says it works closely with customers and customers’ customers to plan capacity and is expanding leading-edge and advanced-packaging facilities in response to strong multi-year demand.
For buyers, this creates an important distinction between having a forecast and having an allocation. A company may expect to sell millions of AI systems, vehicles or devices, but that demand cannot be monetised if the required wafers, advanced packaging or memory components are unavailable during the launch window. Reserving capacity therefore becomes a form of option on future revenue.
The logic also changes supplier relationships. Traditional purchasing tries to maintain competitive tension and avoid overdependence on one vendor. Capacity reservation can pull in the opposite direction because a supplier is more willing to invest when a customer provides demand visibility, capital support or contractual commitment. The customer obtains supply assurance; the supplier obtains confidence that new capacity will be utilised.
That does not mean long-term commitments are always superior. Semiconductor demand is cyclical, technology nodes become obsolete and product roadmaps can shift quickly. A capacity contract that is valuable during shortage can become an expensive liability after a demand correction. The strategic value lies in matching the duration of the commitment to the economic life of the scarcity, not simply signing the longest possible agreement.
AI is turning computing capacity into a balance-sheet decision
Large technology companies illustrate how sharply infrastructure commitments can scale. Microsoft’s 2025 annual report disclosed substantial construction, lease and purchase commitments; it noted that purchase commitments primarily related to datacenters and included open purchase orders and take-or-pay contracts.
The significance is not the accounting line by itself. It is that digital capacity—once perceived as nearly infinitely scalable—now requires years of physical investment and contractual planning. Compute has become entangled with real estate, grid access, cooling, networking and hardware supply. For major AI users, future computing capacity is increasingly something to be financed and reserved, not merely consumed on demand.
This creates a new corporate-finance problem. A company that reserves AI capacity too conservatively may lose product launches, customers or model-training windows. One that reserves too aggressively may be left with high fixed costs if demand, technology or model economics change. The procurement decision therefore begins to resemble a capital-allocation decision: management must compare the cost of unused capacity with the opportunity cost of not having enough.
The better metric may be capacity-adjusted return rather than utilisation alone. A contract can look inefficient if measured only by average utilisation, yet still be rational if it protects high-value peak demand or shortens time to market. Conversely, a fully utilised contract can still be poor economics if the company secured the wrong technology or paid a scarcity premium after the bottleneck had already begun to ease.
Aircraft delivery slots reveal the value of time
Commercial aviation offers another visible example. Airbus reported a backlog of 9,222 commercial aircraft at the end of June 2026, while continuing to ramp production across major programmes. A backlog of that scale means an aircraft order is not simply a purchase decision; it is also a place in a production sequence.
For airlines and lessors, an early delivery position can carry strategic value because fleet growth, replacement schedules and route economics depend on when aircraft arrive. A carrier that waits for traffic growth to become obvious may find that the relevant aircraft are not available for years. The economic asset is therefore partly the machine and partly the timing of access to the machine.
Airbus’s orders and deliveries data through July 2026 shows how production slots are allocated across a very large existing order book. Similar dynamics can appear in engines, maintenance capacity, industrial machinery and other complex manufactured assets where qualification and production constraints limit near-term substitution.
This explains why companies sometimes commit before demand arrives. The alternative is not necessarily to wait and buy later at the same price. It may be to wait and discover that the operational window has moved several years away. When lead time is strategic, time itself becomes part of the acquisition cost.
Why companies are paying before they consume
Contracted capacity can take several financial forms. Some agreements require deposits or prepayments that help suppliers fund expansion. Others contain minimum-volume commitments, take-or-pay clauses or non-cancellable purchase obligations. Cloud infrastructure can involve long lease terms or committed spend. Manufacturing contracts may reserve tooling, clean-room time or production lines. Power arrangements may involve connection charges or long-term energy commitments.
Economically, these structures transfer part of the capacity risk from supplier to customer. The supplier receives a stronger demand signal and can justify capital expenditure. The customer obtains preferential access or certainty. The price of that certainty is reduced optionality.
This has a direct working-capital effect. Prepayments convert cash into future operating capability before revenue is earned. Minimum commitments turn variable costs into more fixed obligations. Deposits can sit on the balance sheet for long periods. If the capacity is denominated in another currency, the company may also acquire foreign-exchange exposure. If the supplier fails, the customer may face both a supply interruption and a credit loss.
Finance teams therefore need to treat capacity contracts as more than procurement documents. They affect liquidity, covenant headroom, scenario planning, cash conversion and, depending on the contract, accounting treatment. A strategically valuable reservation can still weaken financial resilience if it consumes too much cash or concentrates obligations in the same downside scenario.
The hidden risk is over-reserving the future
Scarcity can encourage a dangerous response: reserve everything. When managers have recently experienced shortages, the perceived cost of being under-supplied can dominate the perceived cost of excess capacity. That can lead to overlapping orders, duplicated suppliers and commitments based on demand forecasts that were themselves inflated by fear of scarcity.
The classic bullwhip effect can therefore migrate from physical inventory into capacity contracts. If every customer reserves more than it expects to use, suppliers see an exaggerated demand signal and expand aggressively. When the true demand level emerges, the market can shift from shortage to surplus, leaving customers with take-or-pay obligations and suppliers with underutilised assets.
Technology risk compounds the problem. A reserved server generation can become less attractive if a newer architecture dramatically improves performance per watt. A manufacturing line can lose relevance after a product redesign. A grid reservation can be stranded if planning approval fails. An aircraft delivery position can be less valuable if traffic demand weakens or fleet strategy changes. Contracted capacity is not automatically durable simply because it was scarce when the agreement was signed.
A disciplined capacity strategy needs four tests
| Test | Strategic question |
| Scarcity durability | Is the bottleneck likely to persist for the life of the contract, or is it a short-cycle shortage? |
| Revenue at risk | What revenue, launch date or operational capability is genuinely lost if capacity is unavailable? |
| Substitutability | Can the company switch supplier, technology, location or timing without destroying economics? |
| Balance-sheet cost | How much cash, fixed commitment, counterparty exposure and downside rigidity does the reservation create? |
The first test is scarcity durability. A capacity commitment is most defensible when the bottleneck is structural: highly specialised equipment, limited qualified suppliers, permitting constraints, scarce power connections or production systems that require years to build. It is less defensible when shortage is mainly cyclical and can reverse quickly after industry investment rises.
The second is revenue at risk. Management should quantify what is actually protected by the contract. If a reserved production slot protects a product launch worth hundreds of millions, carrying some unused capacity may be rational. If the contract merely avoids modest spot-price volatility, a long fixed commitment may be excessive.
The third is substitutability. Capacity that is technically interchangeable is less strategic than capacity tied to qualification, location, regulation or proprietary tooling. A firm should know how quickly it could switch supplier, redesign the product, move the workload or alter the operating schedule.
The fourth is balance-sheet cost. Companies should stress-test prepayments, deposits, take-or-pay obligations and cancellation penalties against a demand downturn. Capacity resilience that destroys liquidity is not resilience. The financial structure should allow the company to survive the very scenario in which the reserved capacity becomes least useful.
Capacity contracts can reshape competition
If scarcity persists, contracted access can become a competitive moat. A company with secured chips, power, aircraft, production lines or logistics capacity may be able to grow while peers are constrained. This advantage can be especially meaningful in markets where customers value reliable delivery more than the lowest nominal price.
But there is a policy dimension. When a small number of large buyers lock up scarce capacity, smaller firms may face higher barriers to entry. Concentration can also deepen dependence on a handful of suppliers. Regulators and policymakers may therefore become more interested in transparency, fair access and whether infrastructure bottlenecks are being resolved through investment rather than simply allocated to the strongest balance sheets.
The IEA’s grid work points to one constructive alternative: use existing capacity more efficiently. Flexible connection agreements, grid-enhancing technologies and operational flexibility can unlock capacity without waiting for full new infrastructure. The same principle applies more broadly. Companies should not assume every constraint must be solved by reserving more. Better scheduling, product redesign, shared infrastructure, yield improvement and demand flexibility can reduce the amount of capacity that needs to be contracted in the first place.
What this means for banks and investors
For banks, capacity commitments are becoming relevant to credit analysis because they can sit between operating expense and quasi-capital investment. A borrower with large prepayments or take-or-pay commitments may have less liquidity flexibility than headline cash balances suggest. At the same time, secured capacity can lower operational risk and improve revenue visibility. Credit analysis needs to capture both sides.
For investors, disclosed commitments can reveal how management views future demand. Large capacity reservations may signal confidence, but they should not automatically be read as evidence that demand will materialise. The stronger question is whether the company has a defensible route from reserved capacity to profitable revenue and whether the contract preserves enough flexibility if assumptions change.
Investors may also find value in the suppliers that own scarce capacity: semiconductor foundries, advanced-packaging providers, grid-equipment manufacturers, specialist contract manufacturers, aircraft and engine makers, data-centre operators and infrastructure owners. Yet scarcity can attract capital rapidly. High returns often invite expansion, so the durability of the bottleneck matters more than the existence of a current backlog.
The strategic asset is access, not ownership
The broader shift is that companies are learning to separate ownership from access. A business does not need to own a semiconductor fab, power station, aircraft factory or data centre to benefit from scarce capacity. It may only need a credible contractual claim on the output at the right time.
That claim can be strategically valuable because modern businesses are increasingly assembled from interdependent external systems. The company that can coordinate those systems—production, power, compute, logistics and specialist manufacturing—may be able to execute faster than one with the same capital but weaker access rights.
Contracted capacity is therefore becoming a form of operational capital. It is not capital in the accounting sense, and it should not be valued as though every commitment is an asset. But economically, it can determine whether a company has the capability to turn demand into revenue when the physical world is constrained.
The strongest strategy will not be to contract the most capacity. It will be to identify the few bottlenecks where unavailable capacity would genuinely stop the business, secure enough access to protect those points, and retain flexibility everywhere else. In a world where supply constraints can last longer than product cycles, the ability to obtain the right capacity at the right time may become as important as the ability to finance it.
References
1. International Energy Agency — Electricity 2026: Grids
2. International Energy Agency — Electricity 2026: Executive Summary
3. International Energy Agency — Energy and AI: Executive Summary
5. Microsoft — 2025 Annual Report
6. Airbus — Half-Year 2026 Results, 29 July 2026