OpEx

Operating expenditure (OpEx) is the cost of operations recognized as expense; capital expenditure (CapEx) acquires or improves an asset whose cost may be recognized over its useful life. The IAS 16 overview describes recognition of tangible assets and depreciation. Classification and expense recognition are different from payment timing. Cloud compute is commonly purchased as a service, while servers may qualify as assets; the actual contract and applicable accounting rules determine treatment.

Same capacity, two shapes

The example assumes servers are paid for and available for use in month one, depreciated straight-line over three years with zero residual value. It assumes cloud service is consumed, expensed and paid monthly. Prepayment, payment terms, financing and leases can separate cash flows from these expense schedules.

Under these assumptions, a 600,000 server purchase consumes 600,000 of cash initially. Straight-line depreciation is 600,000/36 ≈ 16,667 per month, so the first year recognizes 200,000 of expense while the remaining carrying amount is 400,000 before other adjustments. A service consumed and paid at 19,500 per month instead produces 234,000 of cash outflow and expense in year one, and 702,000 over three years. The different recognition schedules do not change the amounts actually paid.

Depreciation allocates an asset’s cost over its useful life; it is not a second payment and does not recover the original cash. Conversely, paying for a service in advance does not necessarily mean recognizing the whole payment as expense at once. Payment timing, consumption, and recognition need separate schedules.

These are different cash and expense schedules for assumed comparable capacity. The simple totals omit financing, maintenance, taxes, residual value and changes in required capacity.

Cash flow and reported expense answer different questions

A cash forecast asks whether the organization can fund the payments when due. An expense forecast asks which costs are recognized in the reporting period. Neither measure alone establishes the best infrastructure choice.

Continue the example with annual revenue of 4,000,000 and other operating expenses of 2,900,000. Treat the server depreciation as an operating expense in this simplified comparison. Buying gives 4,000,000 − 2,900,000 − 200,000 = 900,000 of operating profit; consuming the cloud service gives 866,000. Yet year-one cash payments for the capacity are 600,000 and 234,000 respectively. The 34,000 difference in reported profit and the 366,000 difference in capacity cash outflow measure different things.

A focus only on first-year reported profit favors buying in this scenario, while a focus only on first-year cash favors renting. Over the assumed three-year horizon, the undiscounted capacity totals favor buying, before the omitted costs and risks.

An incentive based only on one period’s profit or cash can favor different choices in this example. Show both measures and a comparable service horizon; this is not evidence that accounting incentives always determine a purchasing decision.

Price the flexibility given up by a commitment

Some cloud services let consumption and spending fall when demand falls. A minimum commitment can reduce the price while requiring payment even if the workload disappears. The commercial commitment and the accounting classification are separate questions.

Assume one baseline unit costs 26,000 per month on demand or 17,400 per month under a non-cancelable 36-month commitment. Flat demand costs 936,000 on demand versus 626,400 committed, saving 309,600. If demand halves after twelve months, on demand costs 12 × 26,000 + 24 × 13,000 = 624,000, slightly less than the commitment. Any usage above the committed baseline is assumed to cost the on-demand rate; a larger workload is not included for free.

The commitment buys one baseline unit of capacity. Usage above it is charged at the assumed on-demand rate, so growth is not free. With demand at or below that unit, break-even average usage is 17,400/26,000 ≈ 66.9%; a modest decline can still leave the commitment cheaper. In an illustrative two-outcome forecast, assign a 62% chance to continued use and 38% to cancellation. Continued use means flat usage for all 36 months, while cancellation ends usage after month 18. Survival saves 309,600. Cancellation leaves 313,200 of unused committed payments, offset by 154,800 already saved, for a net disadvantage of 158,400. At a 62% survival probability, expected savings are 131,760.

That expected saving depends on the two assumed demand paths and probability. A commitment reduces flexibility even when purchased services are recognized as operating expenses; it does not automatically become an asset purchase for accounting. Price cancellation, unused capacity and overage separately, particularly for a serving workload with uncertain volume.

References

• IFRS Foundation: IAS 16 Property, Plant and Equipment
• Numerical scenarios, chosen rates and simulations in this entry are illustrative calculations by the author, not measurements reported by these sources.