Infrastructure finance tactics your CFO will love

Typically there is not a one-size-fits-all solution. Here's how to find the best one for you.

Consider this hypothetical scenario: A global enterprise has grown through multiple M&A deals, acquiring numerous data centers along the way. Many of these centers, which are both on and off premises, are aging facilities that require major updates to support current IT operations. They lack sufficient capacity to support both organic growth and the company’s mushrooming data initiatives. 

The company’s managers are looking to consolidate operations and identify the best infrastructure mix to host their applications. What does it take? What's involved in optimizing your data operations and finding the best infrastructure to support that combination?

There are many options to consider, and it can sometimes be difficult to choose the one that matches your budget. The decision is all the more challenging when demand for new technologies is rising rapidly and IT budgets are tight at best. Fifty percent of enterprise IT departments say their budgets are flat or shrinking as organizations look for ways to decrease spending on telecom, staffing, facilities infrastructure, and server hardware, according to the Uptime Institute 2016 Data Center Industry Survey Results.

The range of data center solutions is wide and can include upgrading or retrofitting an existing data center, purchasing or building a new one, leasing space, or outsourcing operations to a service provider. As cloud adoption accelerates, many data center service providers are retrofitting existing facilities, or securing land to build mega data centers and rent space to businesses.

Once you’ve identified the right infrastructure solution for your company, how do you get the finance team on board? 

Rent or buy?

Perhaps the most important financial decision you can make concerns your infrastructure. This is essentially a rent or buy decision. You need to decide whether it’s more economical to operate the entire data center yourself or team up with a colocation partner that can host it for you. A third option is a hybrid solution that blends hosted IT and cloud-based resources.

Start by understanding the financial approach your company prefers. In broad terms, owning or building a data center makes more sense for businesses that have—or can easily obtain—funds for capital expenditures (CapEx) and want to decrease operating expenditures (OpEx). CapEx typically includes the up-front cost of hardware and other infrastructure. OpEx is usually assigned to the ongoing costs of facility staff, maintenance, power, and cooling.

Companies with a lot of capital on hand—financial service firms, for example—usually prefer CapEx investments because they have the liquidity to handle them. Leasing a data center can make sense for businesses that need to conserve CapEx and are OpEx-neutral. Companies that need to borrow funds for an investment will tend to favor OpEx. Startups and newer companies are much more likely to choose OpEx because it more closely matches expenses to revenue and can be more predictable.

Get our report: Navigating IT Transformation - Tales from the front lines

As with most business decisions, timing is important. When you offer up your data center proposal for corporate approval, do so well ahead of the budget approval process. The more input and support the proposal has from both technical and financial stakeholders, the easier it will be to get it funded.  

A company will typically do a financial analysis to determine which strategy is the best fit. Your CFO is more likely to approve IT investment requests that align with the company's financial preferences. Total cost of ownership (TCO) is a calculation widely used to compare the cost of different data center sourcing options. In addition, a net present value (NPV) calculation helps you evaluate the present value of future cash flows for each data center option.

Both calculations will vary greatly with the time period considered. 

You’ll find that more permanent solutions—such as retrofitting an existing data center or building a new one—are typically more expensive options in the short term due to the large amount of initial CapEx required. Colocation services tend to cost more over the long term due to recurring, and usually escalating, annual expenses, which are typically tied to OpEx.

Modular—or prefabricated design, as it is referred to—can accommodate intermediate time horizons because it allows you to add capacity relatively quickly when you need it. Most businesses have varying levels of capacity requirements, so a combination of these solutions is often an optimal approach.

Need for speed

If you want to get your operation up and running quickly, OpEx-centric strategies have a shorter implementation cycle than CapEx-centric strategies because expenses are incurred over time and a large up-front capital investment is not necessary. For example, Netflix owned its own data center infrastructure but found it couldn’t add new capacity fast enough to accommodate its growth, according to a company blog. To accommodate rising demand for its services, Netflix embraced public cloud infrastructure and is now entirely cloud-based.

The speed of approval is also a factor. Often, a large capital investment requires higher levels of corporate approval than a series of ongoing expenses. and companies that go the OpEx route are more likely to stay on par with maintenance and upgrades compared with those that opt for CapEx financing. The CapEx disadvantage is that funds used to build a data center are tied up in that investment and so cannot be used for other projects.

Options such as colocation and managed service are the most sensitive to price increases. Colocation costs are currently low due to high availability, but prices rarely go down once you’re locked into a contract unless you threaten to change vendors. Up-front investment can be a good hedging strategy if you expect costs to increase in the future. The opposite is true if you expect prices to go down.

Similar logic applies to expectations for kilowatt (kW) capacity, the IT equivalent of “space.” Short-term, smaller, and variable capacity needs are typically best met by scalable solutions such as colocation, managed service, and cloud. Larger, longer-term needs will often require more permanent solutions such as retrofit and greenfield construction.

Phasing and timing is the key to securing the best financial results. If you expect your IT architecture to shrink going forward, you should avoid securing more space than you need. If you expect your IT architecture to grow, you should secure space to grow into in the months and years ahead.

Location and control

Greenfield, retrofit, and modular solutions are best for companies that want ownership and control over the data center and its daily operations. Companies that prefer to outsource data center operations will tend to opt for colocation, managed service, and cloud options. In this case, asserting control means ensuring the correct service-level agreement is in place.

Another important factor to consider is your data center’s location. The cost of land, buildings, rent, construction, labor, and service providers can vary greatly by region, so think about where your company’s operations and clients are located. Are you able to leverage trained personnel and existing real estate to maintain market presence in that region?

You also need to consider disaster recovery requirements. Companies that operate more than one data center will typically want to maintain geographic distance between each one. These requirements vary by industry. Location requirements are also driven by your degree of tolerance for latency, meaning the time between the moment a signal is sent and when that signal is received. Finally, network capabilities can vary greatly by region and provider. The regional network may be controlled by a larger supplier and then subcontracted out to local network providers, which can significantly impact performance and latency.

Tax incentives also vary by region, as does availability of renewable resources. More than 20 U.S. states currently offer varying tax incentives for new and existing data center sites as well as single and multi-tenant facilities.

Data centers have been called the factories of the 21st century, in part because more than half of all companies still rely on coal-powered sources to meet the bulk of their energy requirements for IT infrastructure. In recent years, more companies have launched “green” data center initiatives, either installing more energy-efficient equipment or leveraging renewable resources.

In recent years, big cloud service providers like Google and Amazon have made significant progress harnessing renewable resources to power their many data centers. However, those companies make up only a small fraction of the data center energy footprint. Worldwide, small to medium-size data centers consume most of the electricity in this market and therefore represent the biggest opportunity for reduction. The business case for renewable energy is heavily impacted by changing regulatory mandates and incentives, which vary widely by state and country.

In the end there is no one-size-fits-all solution. Most organizations need to run a mix of applications. Some apps should remain in-house within the company’s control. Others can be more effectively hosted by a third party. Most likely, your company needs a combination of solutions to accommodate business cycles and fluctuating capacity, location, and investment strategy requirements.

Financing a data center: Lessons for leaders

  • Infrastructure financing is more than a CapEx vs. OpEx decision.
  • Learn your company’s preferred approach to financing before you pitch the CFO on a new IT infrastructure investment.
  • Most likely, your company needs a combination of infrastructure solutions to accommodate business cycles and fluctuating capacity, location, and investment strategy requirements.

Laura Cunningham will be presenting on this topic at this year’s Data Center World conference on April 4. For more information, click here.

This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.