Cloudera: 52% of Enterprises Seeing Significant Value from AI

Enterprise IT leaders are realizing measurable value with their AI initiatives, according to Cloudera’s 2025 State of Enterprise AI and Data Architecture report released today.

Widespread AI adoption among businesses

Cloudera, in collaboration with Researchscape, surveyed 1,574 IT leaders from various companies worldwide to examine the divide between AI-forward enterprises that treat it as an urgent mandate and those that view artificial intelligence as an unnecessary risk or source of inefficiency.

The study found that 52 percent of respondents reported significant success in achieving measurable value from their AI initiatives, while only 1 percent reported no success at all.

Notably, 96 percent of respondents stated that AI was at least somewhat integrated into a core business function, underscoring the extensive adoption across various industries.

The research also revealed that organizations are utilizing a diverse range of AI models. 

Here are the percentage breakdowns of those who reported using the following types of AI:

• Generative AI is leveraged by 60 percent of respondents 
• 53 percent said they use Deep Learning models 
• Predictive AI made the list for 50 percent of those who answered 
• Supervised Learning (43 percent) and Classification (41 percent) came next in popularity
• Agentic AI was cited by 36 percent
• Regression models were used by 24 percent of those surveyed

Cloudera said that these results reflect growing enterprise confidence in managing newer forms of AI. AI agents, in particular, emerged as a focal point of development, with 83 percent of respondents in Cloudera’s Future of Enterprise AI Agents report stating that investment in AI agents is crucial to gaining a competitive advantage.

Growing trust in enterprise data amid the rise of AI

Data architecture, or how IT leaders are managing data at scale, was another key focus of the study. Nearly one-quarter of respondents (24 percent) stated that they trust their data “significantly more” than they did a year ago, while an additional 41 percent reported trusting their organization’s data “somewhat more.”

This trend highlights positive momentum toward data-driven approaches, despite the hurdles that come with AI adoption and implementation. 

Cloudera also asked IT leaders about the capabilities they most want to have as they leverage their data architectures to fuel AI-powered initiatives. The top responses were:

• 52%: Integrated AI and ML ops tooling
• 51%: Automated data pipeline orchestration
• 44%: Granular data governance
• 41%: Unified data access layer

The survey also explored where companies store their data, with a range of options cited:

• 63% in a private cloud
• 52% in a public cloud
• 42% in a data warehouse
• 38% on-premises mainframe
• 32% on-premises distributed
• 31% in other physical environments
• 25% in a data lake
• 24% in a data lakehouse

Cloudera noted that many companies continue to rely on on-premises storage solutions, even though cloud options were the most popular in the poll. The company highlighted on-prem environments as “highly preferred among IT leaders,” reflecting the trust they retain as AI adoption expands.

Challenges with implementing AI remain

The report also outlined how IT leaders and teams are navigating the various challenges of implementing AI, particularly around the technical and security concerns associated with it.

When asked about the technical limitations they faced in supporting AI initiatives within their data architecture, the top barrier was data integration (37 percent). Other key limitations included storage performance, compute power, and a lack of automation.

Security and compliance also ranked high on the list of concerns, with nearly half (46 percent) of respondents identifying them as top issues when it came to adopting AI. 

Here’s a full breakdown of the top security concerns cited:

• 50 percent cited data leakage during model training
• 48 percent said unauthorized data access
• 43 percent are concerned about insecure third-party AI tools
• 39 percent report they lack visibility or explainability in model outputs
• 35 percent shared concerns over model manipulation or poisoning
• 34 percent cited regulatory non-compliance
• 21 percent reported concerns about hallucinations

These findings indicate that leaders continue to face challenges in adopting and integrating AI into their organizations’ data architectures. Even so, the trend toward adoption is steadily advancing, making broader AI integration in the following year increasingly likely.

Last August, Cloudera acquired Taikun, a platform provider for managing Kubernetes and cloud infrastructure across environments. Read more about the acquisition and how it strengthens Cloudera’s deployment capabilities.