Three essential changes to escape AI stagnation in 2026

The fundamental reason behind this expectation is that AI is very different from past transformational technologies. To realize the full potential of AI, companies will need to change not only how they manage data, but also how they design and invest in IT infrastructure, define job descriptions, design business processes, and track and measure success. These shifts take time.

With that in mind, here are three transformations IT organizations will make in 2026 to prepare for an AI future.

1. AI initiatives will transform IT job descriptions

We now understand that there is a strong connection between high quality data and AI success. This means that efficient use of data is paramount to long-term success in both AI and business. Achieving that success requires changing not only the tools data scientists and engineers use, but also the way they build, manage, and maintain their IT environments.

My colleague Simon Robinson recently highlighted how on-premises data storage infrastructure providers such as Dell Technologies, HPE, Hitachi Vantara, Infinidat, NetApp, and Pure Storage are joining Hammerspace, Vast Data, Weka, and other AI-focused storage players to offer more data management platform-like features in their storage portfolios. Most of these additional features complement existing tools used by data engineering and operations teams. As data storage technology evolves, so do administrator responsibilities.

However, changes in IT architecture responsibilities go beyond data storage. AI efforts are driving the need for greater silicon diversity and the use of heterogeneous computing environments (CPUs, GPUs, etc.). With increasing pressure on both power and cooling, as well as infrastructure budgets, there is a focus on improving utilization based on demand for specific parts of the AI ​​lifecycle, such as preparation, training, and inference.

In addition to storage and compute, networking also plays a key role in AI success. A study by Omdia, a division of Informa TechTarget, found that 50% of organizations considering or investing in private AI initiatives anticipate needing high-performance networking to support those initiatives.

The takeaway from all of this is that infrastructure architects and managers must play an increasingly strategic role when it comes to AI. You need to design, integrate, and manage infrastructure that can optimize data usage and large-scale data pipelines, ensuring efficiency across servers, networking, and storage.

In other words, given the importance of data to AI, data infrastructure can no longer be deployed in silos. As a result, they can no longer be managed as silos. The role of the IT administrator will shift from a project-by-project focus to being a strategic resource for continuous optimization of the larger data management ecosystem.

2. Demand for greater agility and consolidation will change IT infrastructure investment priorities.

Just as IT responsibilities need to evolve, so too must the procurement and design of IT architecture. Most modern IT environments are not ready for enterprise AI at scale. According to Omdia research, when organizations purchase new infrastructure, only 12% actively pursue technologies that provide a consistent IT experience across hybrid cloud locations. This will continue to change if we are to become more data agile.

For many years, the primary focus of new infrastructure deployments has been to meet the needs of the specific application environment they serve. This approach assumes that if data needs to be moved or demand changes, it can be handled by additional tools or services that can be added later.

Given the increasing regularity of data movement and the resulting need for better data governance, the deployment of new infrastructure in the AI ​​era cannot be separated from data pipeline architecture. Therefore, the ability to improve data agility (the movement of data from one location to another) and infrastructure agility (the ability to adjust or extend the core functionality of the infrastructure while maintaining a consistent management experience) becomes a key requirement for new investments. Platforms that can provide these superior agility benefits can provide excellent value for AI initiatives.

Coupled with the importance of agility, integration is becoming a higher priority. Given the complexity, cost, and power burden of infrastructure to support AI initiatives, IT leaders will increasingly consider consolidating the number of systems they manage. Consolidation has long been a byproduct of modernization. Investing in modern, high-performance storage, networking, or server systems provides the benefit of requiring fewer physical systems to achieve the same results.

However, historically, the focus on accelerating adoption and improving time to value has resulted in less-than-ideal post-deployment utilization. Such low initial utilization levels can often be explained as room for future growth. Given the cost and complexity pressures of AI, lower utilization is a luxury that enterprises can no longer afford.

IT infrastructure vendors, especially data storage vendors, have made it much easier to scale their infrastructure on demand and pay for their infrastructure based on usage. These flexible payment options, combined with the performance and agility benefits offered by modern infrastructure options, greatly enhance the benefits of infrastructure modernization for AI initiatives. This should improve integration, reduce cost and complexity, and increase data movement agility.

As the amount of spending directed toward AI initiatives increases, the ability to measure success and quantify value will soon become important, if not already, to determine the success of new AI initiatives.