When it comes to procuring new equipment, capabilities, and software, IT professionals generally have two options:
Procure new equipment as a capital expenditure (CapEx).
Procure “as a service” capabilities as an operating expense (OpEx).
As many companies shift from traditional hardware and software ownership to these cloud or as-a-service models, they need to understand the pros and cons, and navigate the transition. It is not as easy as just flipping a switch or calling your VAR; migrating completely to an OpEx model can take months or even years depending on the size and complexity of the organization.
In this blog, I will cover the following topics:
Define CapEx and OpEx as it pertains to IT purchases in the US.
Give examples or standard use cases for each model
Discuss how a transition from traditional hardware purchases to cloud computing introduces new challenges.
What are capital expenditures (CapEx)?
Capital expenditures (CapEx) refers to the funds a company commits towards buying fixed assets, such as the purchase of servers, storage or network devices in a data center. This equipment traditionally is acquired, paid for immediately, and then depreciated over time. From an accounting perspective, this traditional model places assets on the balance sheet which then declines in value over the life of the asset.
One-time purchases of these physical hardware or software assets are usually intended to benefit the organization for three to five years, or more. Ideally, the depreciation schedule the accounting department uses matches the useful life of the assets. Since different assets may have varying useful lifecycles, it is common for IT departments to use different depreciation schedules based on the type of equipment, and the planned refresh cycle. Adding complication, IT departments may execute hundreds of IT purchases a year, resulting in hundreds of depreciation schedules for the assets on the balance sheet. In the IT world, examples of these major items include:
Servers – usually refreshed every 3-5 years
Storage devices – also refreshed every 3-5 years
Networking gear – core switches and routers often have different refresh cycles than edge devices.
Datacenter equipment – Universal Power Systems (UPS), air conditioners, generators – these large dollar items may have a useful life of 10 years or more
Software and Software development – both can usually be capitalized and there are various depreciation methods used.
CapEx spending has pros and cons from the accounting side. If the asset’s useful life extends beyond a year, then the asset is depreciated anywhere from 5-10 years beyond the purchase date.
As an example, if a company buys a $1 million computer system, it may depreciate it in even amounts over a 5-year period. This means that each year, a $200,000 depreciation expense will be recorded and the asset value on the balance sheet will be reduced by the same amount. While the company’s books may only show a $200,000 expense in the first year, the entire $1 million was likely paid upfront. So, a CapEx expenditure of this size significantly reduces the cash a company has available to invest in its core business.
What are operating expenses (OpEx)?
Operating expenses (OpEx) are the cash outlays that support your day-to-day business. OpEx items are generally consumed within the year they are purchased. Examples include:
Power bills, and other consumables such as office supplies.
Contracted services, such as maintenance agreements, website hosting, and as-a-service subscriptions.
OpEx purchases can be acquired in different methods:
By contract with a predictable flat monthly expense, often by annual commitment. Rent or maintenance contracts are good examples.
Consumables that are immediately used. Paper, snacks, ink cartridges, etc.
Consumption models where you pay for what you use. Electricity and cloud computing bills are good examples. These models are typically very flexible or elastic – fees go up and down dynamically as you consume more or less of the service.
When goods or services are purchased via an OpEx model, all costs are attributed to the operating expense budget and tracked in the P&L (Profit and Loss statement). The entire expense is deducted from the bottom line, instead of being depreciated over several years. By default, these expenses reduce a company’s tax obligation in the year they are made.
Determining CapEx vs OpEx treatment
Though the definitions are usually well-understood, there are grey areas and options. A server can potentially be acquired in multiple ways:
CapEx Purchase: You can pay cash and own the server outright, depreciating it over a pre-determined time frame.
Lease: You may decide to lease the server, typically over a 3 or 5-year term. This method reduces the cash needed at the time of acquisition but usually doesn’t give the flexibility of a consumption model. Furthermore, accounting treatment may still require the lease obligation to be on the balance sheet.
OpEx: Signing a contract with a managed service provider or cloud computing provider. In this model, a company doesn’t own the asset, so it doesn’t appear on the balance sheet. You pay for what you use as an operating expense.
Having the choice between CapEx and OpEx for acquiring new IT capabilities isn’t a novel development. These options have been with us in various forms for a long time. The cloud has dramatically changed the landscape so that most IT products and services can easily be bought via an OpEx model.
Comparing CapEx vs OpEx for IT
Outside of the tax and payment treatments, there are several advantages and disadvantages to procuring major IT capabilities as either CapEx or OpEx items. For example, when comparing the different methods for purchasing a server, there are several factors to consider.
Upfront costs: For a capital purchase, all money is typically paid up-front.
Approval process: CapEx and OpEx expenditures affect different budgets, so OpEx purchases are usually faster to implement.
Implementation and Support: Purchasing products as CapEx items may also require you to buy several other supporting capabilities, like maintenance, software, or additional data center infrastructure.
Operational Control: In an OpEx model some or all support may shift to an outside provider. This may be a huge advantage, or be seen as the opposite, depending on the business and existing IT support team.
Planning: buying a product as CapEx requires lots of planning, as you are likely going to use this product for years. Buying products in an OpEx or as-a-service model allows for pay-as-you-go contracts where forecasting usage is much less important.
Scalability: Doubling server capacity in an OpEx model means you need to make additional OpEx purchases. Scaling either up or down in the cloud is usually much easier.
New Challenges by Moving to the Cloud
Companies that favor OpEx models see the cloud as the perfect solution. They love the model whereby you pay only for what you use and pay as you use it. However, transitioning to OpEx is often a multi-year process. As stated earlier, you can’t just flip a switch to make the transition. Three common barriers that are often overlooked are Long-term contracts, refresh cycles, and lack of cloud talent.
Long-term Contracts: Enterprises usually have a multitude of contracts that have terms of 1-year or more. These might be OEM maintenance agreements, lease agreements, colocation contracts, or outsourcing agreements. Planning a migration that coordinates with contract end dates is sometimes considered, but rarely easy to execute due to the uncertain time it takes to perform a migration. In addition, enterprises may have hundreds of these contracts, all with different end dates.
Refresh Cycles: Companies may refresh critical IT infrastructure every 2 to 5 years. The challenge is that the value of IT equipment is rarely as high as the book value on a company balance sheet. Most are aware of the phenomenon where a car may lose 10% of its value the minute it is driven off the lot. IT equipment is exponentially more challenging. Restrictive software agreements by OEMs often make IT equipment lose over half its value when “driven off the lot”. Companies often talk about “sweating out” their assets by letting them get older before doing a cloud migration. This allows them to take less of a write-off but is really just a band-aid to a problem. This ignores the business reasons for doing a migration in the first place. Letting refresh cycles drive your business strategy is rarely a winning formula.
Lack of Cloud Talent: In addition to the challenges of owning legacy hardware, the skill sets for managing a cloud environment are often very different. Companies may not have any of the expertise to either migrate applications themselves or manage the applications post-migration. Building these skillsets internally takes time, energy, and a dedicated focus.
Ideally, a company looking to move to the cloud and to an OpEx model can take interim steps to achieve these goals. The main thing is to avoid continuing old practices that restrict flexibility: don’t refresh hardware, don’t renew any OEM maintenance contracts, and if new IT equipment is needed in the short-term, rent instead of buy. And build an internal team of cloud champions or a cloud center of excellence.
Conclusion
There are many advantages to moving to an OpEx model, in particular in moving to cloud computing. But changing to this model takes careful consideration for what will be left behind. Datacenters, IT hardware, Software licenses, OEM maintenance agreements, and even old hardware skill sets may no longer be needed. Smoothly transitioning away may be the key to a successful migration.
