The potential of generative AI continues to take hold across organizations globally: According to Enterprise Strategy Group, over half of respondents (53%) to a recent survey said that their most significant AI investment over the next 12 months would be in generative AI (GenAI).¹ For many organizations, their GenAI implementation will be hybrid, typically spanning a combination of on-premises data with public cloud-based large language models (LLMs).

However, the devil is in the details, and for many organizations there’s a considerable gap between their high levels of anticipation around AI, and the realities of an effective implementation. For example, Enterprise Strategy Group research found that 62% of organizations had experienced ‘extensive’ or ‘moderate’ challenges moving AI models from development into production.²

Enterprise challenges around implementing AI are broad and vary widely. But while much of the initial focus was focused on the compute environment, many challenges are now emerging at the data level.

Data is the lifeblood of any AI initiative—the one aspect that can make the difference between success and failure. Getting the data aspect right at scale presents numerous substantial challenges across the broader data environment. A recent Enterprise Strategy Group research study found that data management and/or data quality issues were the second most frequently cited challenge associated with implementing AI, behind overall cost issues (see Figure 1).³ Concerns over data privacy, protecting intellectual property, and security were also frequently cited, along with integration issues and the need to modernize infrastructure.

Many organizations on their AI journeys are, therefore, concluding that modernizing the infrastructure—right down to the storage environment—might be a necessary step to fully take advantage of AI, especially as they move from model training to the inference phase.

Figure 1. Top Challenges With Implementing AI

Source: Enterprise Strategy Group, now part of Omdia

^1. Source: Enterprise Strategy Group Research Report, 2025 Technology Spending Intentions Survey, December 2024.

^2. Source: Enterprise Strategy Group Research Report, Navigating Build-versus-buy Dynamics for Enterprise-ready AI, January 2025.

^3. Ibid.