The Hidden AI Tax: IDC Research Reveals Nearly All Organizations Lose Cost Control When Deploying GenAI and Agentic Workflows at Scale
71% of respondents admitted that they have little to no control over where AI implementation costs are coming from
December 9, 2025 — BOSTON — A new IDC survey sponsored by DataRobot, the agentic workforce platform, found that 96% of organizations deploying GenAI and 92% implementing agentic AI reported costs were higher or much higher than expected. Adding to the challenge, 71% admitted they have little to no control over where those costs are coming from.
This financial blind spot is creating a winner-take-all market for those early movers who embed governance and cost visibility from the outset. In fact, organizations that have been experimenting with agents expect to run over 100 agents within two years, while those in the pilot phase only expect to deploy ~25 in the same timeframe. That deployment gap represents a growing competitive chasm that could leave organizations in the pilot phase stuck with heavy implementation costs and no ROI as early movers gain market share.
The AI Cost Crisis: Where the Money is Really Going
The report, which surveyed hundreds of decision-makers, found that nearly every enterprise organization is underestimating the cost of implementing AI across the entire lifecycle. While 79% of enterprises have adopted AI agents in some form, the road from pilot to production is proving to be littered with financial landmines. The survey uncovered inference as a common cost challenge, but token consumption and hallucination remediation have emerged as top unexpected costs when deploying agents.
The math is brutal: more agents mean more vendors, more people, and costs spiraling in ways traditional CIOs simply can’t track.
Operational Cracks Widen at Scale
At scale, cracks appear at every stage of AI deployment for enterprises. Organizations that have deployed GenAI across over 75% of their departments rely on tools from six vendors on average. Additionally, nearly half (48%) of their IT workforce is consumed by GenAI and agentic AI work, up from one-third at organizations with limited deployments — and much of it can be associated with the cost of having to manage and stitch multiple tools and vendors together.
“The 96% cost overrun rate isn’t just a budget problem, it’s an early warning sign that most enterprises are scaling blindly,” said Venky Veeraraghavan, Chief Product Officer at DataRobot. “The data also revealed a dangerous outlook: organizations that can’t control costs at 10 agents certainly won’t be able to at 100. Meanwhile, early movers that have invested in a foundation of governance, unified tooling, and cost visibility from day one are pulling away fast. This gap will define market leaders for the next decade.”
The data is clear: what separates those seeing real business impact from organizations that can’t move past the pilot stage is not how quickly they experiment. In contrast with those stuck in the pilot phase, early movers have implemented enterprise-grade controls from the get-go.
This means that they have embedded governance from the beginning, have leaned on unified platforms to limit tool sprawl, and have built-in cost visibility from day one to limit being blindsided by unexpected costs. In fact, the survey found that organizations that have had GenAI in production for a year or more are 2x as likely to cite the use of an end-to-end enterprise AI platform as critical in reaching GenAI and agentic AI goals.
To download a complimentary copy of the IDC report, “The Agentic Workforce: From Experimentation to Scalable Impact,” visit https://www.datarobot.com/resources/a-strategic-approach-to-scaling-generative-and-agentic-ai/.
Methodology
This IDC InfoBrief was commissioned by DataRobot to explore AI adoption, cost dynamics, and vendor strategies. IDC conducted this research using an online survey distributed among n=318 senior decision makers at organizations with over 1,000 employees across the United States, the United Kingdom, and Ireland. Data was collected in July 2025.
About DataRobot
DataRobot empowers AI teams to deliver the agentic workforce of the future. Our platform enables organizations to create and scale AI agents that integrate directly with business processes—driving efficiency, transforming operations, and delivering real results. With built-in governance and safeguards, we help enterprises deploy AI securely and confidently. For more information, visit our website and connect with us on LinkedIn.
Contact: press@datarobot.com