Authors: Rick Moore, Senior Director, Business Development and Rick Galietta, Director, Enterprise

IT leaders face growing pressure to produce solid benefits from a data-driven organization. That pressure often comes from highly complex data and analytics stacks across the enterprise.

Savvy leaders in IT know their biggest challenges lie in:

An ever-shifting data regulation landscape
An explosion of an endless number of endpoints, applications, service providers (SPs), and users
The need to bring results from the most data-dependent parts of the business

These challenges have often led to more complexity than necessary, and disorganization between cloud service providers, users, partners, and legacy applications.

Add to this scenario internal pressure to integrate Artificial Intelligence (AI) applications into current IT solutions. Plus, AI presents additional challenges since it requires a unique environment that handles high-density workloads and high-performance compute. This complexity makes the path to effective IT budget management ‌less clear.

This article serves to provide a high-level view and considerations for re-architecting data infrastructure, Hybrid IT models, and IT financial management solutions. A clear strategy that enables the deployment of AI and hybrid clouds will require a more flexible, comprehensive plan for managing capital expenditures (CapEx) and operational expenditures (OpEx).

Best practices in AI-enabled data architecture

Capital expenditure, or CapEx, refers to one-time investments needed to acquire assets. Operational expenditures, or OpEx, is money spent on an on-going basis for a business to run operations.

For years, IT leaders viewed CapEx and OpEx through a cloud-computing lens, defining CapEx as the purchase of servers and hardware while attaching OpEx to buying capacity and scaling up or down as needed.

According to 451 Research, “nearly three in five organizations (53%) increasing budgets in 2024 are looking to spend on generative AI.”¹ So as algorithms, training models, and generative AI applications gain more traction in enterprises, IT leaders will look to define, manage, and budget for CapEx and OpEx in more innovative ways.

One best practice is to consider OpEx as a tool, giving enterprises the flexibility of scaling up or down depending on demand. OpEx offers a competitive advantage for those with the foresight to leverage Hybrid IT, by focusing on core competencies rather than hiring or relying on in-house resources to manage a hybrid cloud environment.

Conversely, enterprises must invest in the right hardware as strategies around AI and high-performance computing develop and integrate into day-to-day workflows. This attention to CapEx budgeting and investment would serve companies better if it remained in-house and under the direction of experienced IT leadership.

This leads us to take a deeper dive into modern use cases for CapEx and OpEx.

Leverage a combined CapEx and OpEx model to develop an optimal Hybrid IT environment

When we look at IT modernization from a CapEx vs. OpEx perspective, we find advantages to both. Ultimately, the true strength of the CapEx/OpEx hybrid investment model lies in creating a Hybrid IT environment.

Modern use cases for CapEx include:

Pooling resources in legacy, private cloud / on-prem data centers
Hardware refresh
Enterprise software investment
Patents and exclusive usage rights to intellectual property

The advent of AI has forced enterprises to make tough CapEx decisions around legacy data centers, throwing depreciation timelines into disarray. With the introduction of every new semiconductor designed for AI workloads, the value of legacy servers can plummet since they sometimes can’t handle the increased capacity AI semiconductors provide.

It's becoming more and more unsustainable to keep pace with the quickly evolving power density needs and the cooling requirements to cool those power needs.

The exciting part of emerging IT modernization is that OpEx provides for thoughtful, cautious evaluation of AI applications within an enterprise. For instance, savvy, budget-responsible organizations are using colocation as the bridge between legacy architectures such as on-prem and flexible architectures of OpEx such as cloud computing. Again, keeping pace with the fast-evolving GPU architectures supporting Gen AI and inference models is difficult for most enterprises.

Now let’s shift to exploring how hybrid cloud positions organizations to adjust to emerging risks and complex cloud pricing structures.

Hybrid cloud: How OpEx creates agility in IT modernization

The need for both an agile public cloud and secure private cloud environments isn’t isolated to highly-regulated industries. Thanks to AI, more data governance regulations put more restrictions on where enterprises can house certain forms of data. Here, we’ll review the benefits and areas of complexity of hybrid cloud:

Benefits of Hybrid Cloud

Competitive advantage of an IT modernization-ready architecture
Customized solutions for complex challenges
Flexibility to house workloads in the most effective (cost and operational) environment
Agility to scale up or down while maintaining stability and security on the enterprise’s most sensitive workloads

Areas of Complexity with Hybrid Cloud:

Streamlined allocation of workloads
Coordination between environments
Development of data silos
Hiring ‌dedicated headcount to manage complex cost structure of hybrid cloud environments

When done with intention, hybrid cloud makes room for global enterprises to move funds between systems, devices, software, and services as enterprise needs evolve.

From an IT budgeting perspective, analyst and research groups such as Gartner predict an 8% growth in IT spending in 2024. Much of that spending, Gartner projects, will be in IT services and software.

To plan, predict, and test IT budget scenarios, enterprises also turn to colocation to benefit from the best of hybrid cloud.

Data center modernization and colocation

In successful IT modernization, modern IT operations master technology workflows and workloads that serve their primary business while releasing day-to-day management of other non-core aspects to specialists. This gives companies reduced CapEx exposure while freeing up resources for strategic use of OpEx. Here’s an example:

An example of coordinated growth through colocation data centers comes to us from one of our brewery customers. The business made the decision to shut down their on-premises environment and move to the cloud. As with any large operation, it had many moving parts and some data needed to remain on physical hardware under the brewery’s control.

The business decided the best option was to use a Digital Realty data center to ease the transition between on-premises/private cloud and public cloud. For instance, while migrating to the cloud, the brewery wanted to identify which workloads were best suited for a public cloud environment. Through their colocation data center space, it was able to place some workloads on public cloud and pull them back to brewery-owned hardware once they concluded their testing.

This saved the brewery the steep costs of moving the wrong workloads to the wrong environment. It also gave the brewery a large advantage over competitors in the market because of the hybrid cloud agility this one strategy provided.

An optimized methodology to enable agile IT strategies

In today’s article, we touched on combining CapEx and OpEx investments while dispersing workloads across hybrid cloud environments and across global locations, all while maintaining control over data, cybersecurity, and interconnections.

To make all of these moving parts come together, Digital Realty offers Pervasive Datacenter Architecture (PDx^®). It’s an optimized methodology that codifies hundreds of production deployment combinations into an interconnected, decentralized, and agile data architecture.

The PDx^® Blueprint features a private workshop, a data architecture design guide, and methodology for implementing a streamlined data landscape that provides your organization with more agility by unlocking more power in your data. Find out more by emailing workshop@us.digitalrealty.com with “workshop” in the subject line.

^¹451 Research, Data Insight: Generative AI may spark IT spending turnaround in 2024, November, 2023.

ServiceFabric Service Directory