Embracing a Hybrid IT Environment: Three Pillars of a Successful Edge Computing Strategy

By: American Tower

To meet escalating compute requirements, enterprises are increasingly moving beyond a centralized data center model and adopting hybrid IT environments. This approach capitalizes on edge computing to scale reliably, optimize resources, and unlock new revenue opportunities.

Highly centralized data centers aren’t always the most effective option for some applications because their users are hundreds of miles away and may not meet requirements for low latency.

Why Artificial Intelligence (AI) is Driving Edge Adoption

Edge computing addresses this challenge by processing data closer to its source, reducing latency and enabling rapid decision-making. This is critical for AI, automation, and other applications where milliseconds matter.

AI workloads, for example, are often divided into two main categories: training and inference. AI training is the process of “teaching” a large dataset, such as a Large Language Model (LLM), to recognize specific patterns or how to perform complex tasks. AI training requires significant compute power, which is why this process typically occurs in hyperscale or regional data centers.

Inferencing is the execution phase of AI, applying trained models to make predictions or decisions based on new or unseen data (i.e., responding to user queries). Inferenced models typically require lower latency and less computing power than other use cases—making them ideal for edge environments. According to Boston Consulting Group research, inference workloads are projected to increase by 122% between 2023 and 2028.¹

How to Evaluate Edge Capabilities for Your Enterprise Needs

As organizations shift toward distributed architecture models, consider the following factors as you evaluate how edge compute can support evolving business requirements.

1. Match Workload to Infrastructure

Not every workload needs to run from a centralized data center, as misplacement can cause higher costs, degraded performance, and create other challenges. Consider the following guidelines to make optimal network architecture decisions:

• Review the latency tolerance of an application. How sensitive is your application to response times?
• Think about your compute footprint, including power and density.
• Evaluate your network connectivity and backhaul requirements. What bandwidth or redundancy does it need?
• Analyze the cost implications. How do egress fees and energy usage affect your total cost of ownership?

Organizations that map these factors across the cloud-to-edge spectrum can gain a clearer understanding of workload optimization. By placing workloads effectively across a network, companies can avoid overprovisioning or underutilizing infrastructure.

2. Gain Scalability and Flexibility With Hybrid Cloud

Cloud architecture has been a critical enabler of scalability for years. However, drawbacks to hosting 100% of workloads in the public cloud include latency, data sovereignty, and unpredictable egress costs.

Hybrid models address these challenges by integrating an enterprise’s internal IT resources with a third-party cloud provider’s infrastructure and services. This approach allows organizations to store data and run applications across multiple environments, improving performance while controlling expenses.

As organizations extend their business applications to the edge, cloud access becomes increasingly important. Cloud exchange platforms enable businesses to connect their own IT infrastructure directly to cloud service providers.

A hybrid edge-to-cloud model can balance the strengths of centralized and distributed environments, providing elasticity while maintaining control over performance and costs.

3. Future-Ready Your Business for Low-Latency Compute

Technology change is constant, and what’s cutting-edge today could be outdated tomorrow. Here are a few best practices for capitalizing on your edge investments:

• Address Real Estate Needs Proactively. Land and power access for data center development was once an afterthought. Today, it is often the gating factor that delays or prevents enterprises from introducing new capabilities required for growth. Partner with experts that understand land and infrastructure requirements for data center development.
• Plan for Next-Generation Compute. AI and high-performance computing workloads are driving new requirements for power and cooling, such as liquid cooling. Identify vendors capable of accommodating a wide range of use cases.

American Tower’s Edge Capabilities and Real Estate Portfolio

American Tower is uniquely positioned to address customers’ data center and edge computing needs. We are one of the largest publicly held Real Estate Investment Trusts (REITs) and a leading independent owner, operator, and developer of digital infrastructure.

In 2021, we acquired CoreSite, one of the largest U.S. data center companies. By combining the strengths of our distributed real estate footprint with CoreSite’s interconnection platform, we are able to deliver local data center infrastructure that enables edge computing.

Extend Your Infrastructure to the Edge

Selecting edge capabilities is part of a broader strategic plan to architect a distributed ecosystem that aligns with your workload demands, scales economically, and has the flexibility to address future advancements.

Choosing the right infrastructure for the right application isn’t just a technical decision. It’s a strategic one that impacts your organization’s agility and competitiveness.

Learn More About American Tower’s Data Center Portfolio

¹https://www.bcg.com/publications/2025/breaking-barriers-data-center-growth