In 2025, the companies that are truly realizing value from AI are those that have prepared the groundwork, modernized their data infrastructure, ensured information security, and established strong governance. This is demonstrated by the new global study on enterprise AI maturity prepared by IDC and promoted by NetApp, which analyzes how AI adoption has evolved around the world and what factors are making the difference between success and stagnation.
The new era of enterprise AI, from testing to impact
If 2024 was the year of experimentation, 2025 is the year of consolidation and return on investment. The IDC report shows that organizations have gone from wondering how to implement AI to questioning how to scale it and obtain measurable results. The most mature companies, called “AI masters,” are already seeing notable improvements in their performance, with a 24.1% increase in revenue and a 25.4% cost savings compared to companies that are still in the early stages of adoption.
This leap does not depend so much on the use of generative models or large investments in software, but on the solidity of the foundation on which the projects are built. Data quality, the agility of cloud architectures, and the integration between security and governance are now the true differentiators.
In 2025, the companies that are truly realizing value from AI are those that have prepared the groundwork, modernized their data infrastructure, ensured information security, and established strong governance.
“AI is no longer about demonstrating its viability, but about demonstrating its value”
Syam Nair, Chief Product Officer at NetApp, sums up the situation clearly, “AI is no longer about proving its viability, it’s about proving its value. IDC’s latest research shows that the real differentiators are data and infrastructure preparation: companies that focus on data quality and creating modern, intelligent, cloud-based, scalable and adaptable architectures are the ones that are converting AI in true business impact.”
For Nair, the current challenge is not technical, but strategic. AI needs to be supported by an intelligent data infrastructure, capable of managing massive volumes of information safely and efficiently, “all organizations need an intelligent data infrastructure to be successful in the age of AI.”
Infrastructure, the great bottleneck
Although progress is visible, the study reveals that infrastructure remains an obstacle for many companies. 84% of organizations admit that their storage is not fully optimized for AI, although this figure has improved compared to the previous year. Only a third of companies say they have completed a complete review of their storage systems, reflecting that there is still a long way to go to reach the necessary maturity.
On the other hand, security has become a priority, and more than 60% of the most advanced companies in AI have increased their cybersecurity budgets specifically to protect artificial intelligence initiatives. In contrast, just 16% of less mature organizations have taken similar measures.
Agentic AI, a frontier that requires preparation
The NetApp and IDC study also introduces a new key concept, “agent AI”, that is, autonomous systems capable of acting and interacting in real time with other business applications. To take advantage of this type of AI, companies need a secure, structured and accessible database.
Organizations with greater technological maturity are already exploring this new frontier, while those less prepared risk being left behind, trapped in isolated content generation or partial automation projects.
From experiment to production
Dave Pearson, VP of Research at IDC, sums it up: “Enterprises that modernize their data pipelines, governance frameworks, security approaches, and storage architectures are the ones that turn AI pilot projects into production-grade applications that deliver the best measurable business outcomes.”
According to IDC, the difference between media noise and the true impact of AI lies in the discipline with which companies manage their data. The “AI masters” have understood that it is not about implementing more models, but about improving the infrastructure that supports them. These companies are migrating towards intelligent, automated and context-sensitive cloud architectures, where security and scalability are basic pillars.
