Fivetran, a global leader in data movement, has released its 2025 report, 2025 and Beyond: How a Strong Data Foundation Fuels Enterprise Success and AI-Driven Innovation, highlighting a critical gap in enterprise AI readiness. Despite 89% of technology leaders planning to use proprietary data to train large language models (LLMs) in 2025, only 49% believe their current data architecture can support AI demands, underscoring challenges in integration, scalability, and compliance.
89% of tech leaders plan to train LLMs with proprietary data in 2025.
Only 49% believe their data architecture can handle AI workloads.
68% rely on 50+ data sources for decision-making, complicating integration.
64% of CIOs delay innovation due to compliance concerns.
50% plan to invest $500,000+ in data integration next year.
48% expect increased responsibility for data privacy and compliance.
Enterprises are rapidly adopting AI, with 89% planning to leverage proprietary data for LLMs in 2025. However, Fivetran’s survey of 500 U.S.-based C-suite tech leaders reveals that only 49% trust their data infrastructure to meet AI’s demands. Key obstacles include integration complexity (36%), scalability limitations (34%), and security and compliance risks (33%). “If you don’t centralize your data, you can’t see what’s really happening in your business,” said George Fraser, CEO of Fivetran.
With 68% of respondents relying on 50 or more data sources, integration complexity is a significant barrier. Additionally, 64% of CIOs report delaying innovation to address compliance issues, highlighting the tension between AI ambition and regulatory requirements. Legacy systems, lack of automation, and talent shortages further exacerbate these challenges, with 45% citing insufficient self-service capabilities and 41% noting data team skill gaps.
To address these issues, 50% of executives plan to invest $500,000 or more in data integration in 2025, focusing on automation, real-time data access, and governance. Companies like Sedgwick are already seeing benefits. “Now that our data is centralized and governed with Fivetran, we’re building AI tools that surface real-time insights for our claims examiners,” said Adam Fisher, Chief Data Officer at Sedgwick. These tools provide adjusters with historical claim data, enhancing decision-making.
The report also notes a shift in technology leadership, with 48% of respondents expecting greater responsibility for data privacy and compliance, and 45% anticipating a larger role in enterprise-wide data strategy. As AI moves from pilots to production, a robust data foundation is critical to ensure scalability and compliance while driving innovation.
Fivetran’s findings emphasize the urgency for enterprises to modernize data infrastructure to support AI initiatives. While enthusiasm for AI is high, success hinges on overcoming integration and compliance challenges through automated, scalable solutions, enabling organizations to harness AI’s full potential in 2025 and beyond.
Fivetran, the global leader in data movement, is trusted by companies like OpenAI, LVMH, Pfizer, Verizon, and Spotify to centralize data from SaaS applications, databases, files, and other sources into cloud destinations, including data lakes. With high-performance pipelines, seamless interoperability, and enterprise-grade security, Fivetran empowers organizations to modernize their data infrastructure, power analytics and AI, ensure compliance, and achieve transformative business outcomes.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.