Industry executives and experts share their predictions for 2026. Read them in this 18th annual VMblog.com series exclusive.
By Derek Gillespie, CEO, 365 Data Centers
To stay ahead in 2026, enterprises must stop treating AI as a side project and start treating it as core infrastructure. That means designing networks, data centers, security, and operations around AI-driven analytics, automation, and decision-making – not bolting on AI capabilities after the fact.
AI is no longer just an experiment. It is rapidly becoming the logical layer of the entire digital enterprise. The organizations that recognize this – and hardwire AI into their infrastructure strategy – will be the ones that move faster, operate leaner, and compete more effectively.
From AI “Projects” to AI as a Utility
Over the last few years, AI has moved from proof of concept to production at remarkable speed. Recent enterprise studies show that around 87% of large enterprises now report having implemented some form of AI solution, according to the Second Talent tech agency. With generative AI specifically, over 85% of enterprise leaders say they use these tools at least weekly, and three out of four report positive returns on their investments.
In other words, AI has clearly crossed the line from experimental to mainstream.
Yet, in many organizations, AI is still being handled as a set of disconnected initiatives: a chatbot here, a pilot analytics tool there, an automation script in one department. This “project-by-project” approach made sense in 2023. It does not make sense for 2026.
If AI is going to support critical business decisions, customer experiences, and operational processes, it needs the same rigor and planning that have always been applied to core infrastructure: resilience, scalability, performance, governance, and security. That shift in mindset is what will redefine digital infrastructure in the year ahead.
Why the Underlying Infrastructure Has to Change
AI doesn’t just change software. It changes the demands placed on the underlying infrastructure. Three pressures stand out:
First is scale and performance. AI workloads are both data hungry and compute intensive. As AI models and datasets grow, traditional centralized architectures struggle with latency and bandwidth constraints. This is one reason the edge data center market is projected to grow from tens of billions of dollars today to more than $80 billion to $100 billion by the early 2030s, driven largely by low-latency, real-time AI and IoT applications, Precedence Research predicts.
Second is energy and sustainability. Data centers already account for roughly 1.5-2% of global electricity consumption, and multiple forecasts suggest that figure could roughly double by 2030, with AI cited as a primary driver, according to Strategic Energy Europe. AI-focused infrastructure is rapidly becoming one of the most energy-intensive parts of the digital economy. That means efficiency in, where, and how AI runs is now a board-level concern, not just an IT metric.
Third is the placement of intelligence. Many enterprises are realizing that sending every request to massive, centralized AI models is neither cost-effective nor operationally efficient. This is why surveys show that a growing share of organizations are moving toward edge AI for real-time processing and privacy, with nearly three-quarters of AI-forward enterprises planning greater use of edge and decentralized models in the next few years, according to analysis by BIS Research.
These pressures all point to the same conclusion: AI cannot simply “sit on top of” the existing infrastructure. The infrastructure itself has to evolve to make AI practical, sustainable, and reliable.
Edge, Cloud, and Core: AI as an Optimization Layer
The defining pattern for 2026 and beyond will be hybrid and distributed AI, not just hybrid cloud. Instead of assuming that all AI must run in a hyperscale cloud or a single centralized data center, leading architectures are shifting toward a layered model.
At the edge, small and efficient models run close to where data is generated – in local data centers, regional facilities, or edge sites near users and devices. These models can handle the bulk of routine inference, such as classification, routing, anomaly detection, and basic automation. This reduces latency and bandwidth usage and keeps sensitive data closer to its source.
In regional or enterprise data centers, more powerful AI services aggregate insights from multiple edge locations, support organization-wide analytics, and run more complex workflows. This layer also becomes the governance hub for AI, handling logging, auditing, and policy enforcing.
In centralized or hyperscale environments, large foundation models and highly specialized workloads are invoked selectively for tasks that genuinely require heavy-duty compute or broad knowledge, such as complex language understanding or very large-scale simulations.
The net effect is that AI becomes an optimization layer across the entire infrastructure. The network, compute, and storage topology is designed around intelligent decision-making: What should be processed locally, what should be aggregated regionally, and what truly needs to be escalated to a large centralized model.
This approach directly addresses the cost, latency, and energy challenges described above. It also aligns with where the market is headed. Analysts estimate that a rapidly growing share of enterprise-generated data – approaching half now – will be processed at or near the edge, rather than in centralized locations, specifically to support use cases like real-time analytics and AI-driven automation, as reported by market research company Technavio.
In 2026, the infrastructure conversation is no longer “cloud versus data center.” It is: “How do we design a fabric – spanning edge, core, and cloud – that delivers AI where it makes the most sense?”
What Enterprises Should Prioritize in 2026
If AI is to be treated as a core infrastructure requirement, enterprises should focus on a few practical priorities in the coming year:
- Architect for AI, don’t just host it. Instead of asking, “Where can we run this AI tool?” start with “What should our decision and automation fabric look like?” That means mapping critical data flows, key decision points in operations, and latency-sensitive processes, then deciding which parts of that fabric belong at the edge, in regional data centers, and in the cloud.
- Move beyond isolated pilots. The data is clear: Organizations with a coherent, company-wide AI strategy see far better outcomes than those running disconnected experiments. One recent enterprise survey found that companies with a formal AI strategy report more than twice the success rate of those without one, reports LLM company WRITER. In practice, that means aligning IT, security, data teams, and business owners around a shared roadmap that includes infrastructure.
- Design for observability, governance, and risk from day one. As AI spreads into more workflows, questions of data quality, bias, and model behavior move from theoretical to operational. Many enterprises now cite data quality and governance as their top AI challenge, even more than talent or tools. Infrastructure must support robust logging, monitoring, policy enforcing, and auditing for AI workloads just as it does for traditional systems, according to McKinsey’s “The State of AI 2025” survey.
- Prioritize efficiency and sustainability for AI workloads. With data center electricity consumption projected to potentially double by 2030 and AI a major driver of that growth, efficiency is no longer optional, as Deloitte analyzed last year. Placing the right workloads at the right layer, using smaller models where possible, and optimizing cooling, power, and utilization are all now strategic decisions.
- Think “AI-native” operations, not add-on automation. The end goal is not just to sprinkle AI across existing processes. It is to redesign operations so that AI is embedded in how work gets done. Surveys already show that enterprises using AI at scale report double-digit gains in efficiency and cost reduction, especially in process automation and analytics, according to a report published by Wharton’s Human-AI Research.
The Infrastructure Gap Will Define the Competitive Gap
In 2026, most large enterprises will have some form of AI. The differentiator will not be whether AI exists, but how deeply it is integrated into the infrastructure that runs the business.
Enterprises that continue to treat AI as a collection of tools will wrestle with silos, escalating costs, and inconsistent results. Those that redesign their digital infrastructure – spanning edge, data center, and cloud – around AI-powered analytics, automation, and decision-making will be able to move faster, operate more efficiently, and respond to change with far greater agility.
AI is no longer a feature. It is becoming the operating layer of modern infrastructure. The sooner organizations start architecting for that reality, the better positioned they will be for the decade ahead.
##
ABOUT THE AUTHOR
Derek Gillespie is the Chief Executive Officer at 365 Data Centers. Prior to becoming CEO, he served as the company’s Chief Revenue Office, responsible for the sales, marketing, and the commercial go-to-market activities of the business. Derek comes to 365 with more than 25 years of experience in the Infrastructure as a Service (IaaS) space and has been recognized for his leadership and business acumen on a global scale.
Prior to joining 365 Data Centers, Derek was the Chief Sales Officer at Zayo, which became one of the largest privately held operators of global communications infrastructure worldwide during his nine-year tenure. Prior to that, he served as the CRO at Cosentry, the leading provider of public, private, and hybrid IaaS companies in the Midwest. Before joining Cosentry, he spent 11 years at Savvis, serving in various leadership functions of the business, transforming the sales organization into one of the leading providers of colocation, managing hosting and cloud services in the United States, Europe, and Asia.
In 2023, Derek was recognized by Clari as one of the Top 20 CSO’s globally. Prior to that, he received similar recognition by Collective and the Modern Sale as one of the Top 100 CRO’s in 2021.
Derek is an alumnus of the U.S. Naval Academy, where he majored in history and political science.
