# Cloudera Survey Finds Governance and Rising Token Use Constrain AI Scale By Simon Yoon Canonical URL: https://www.tokenpost.com/news/technology/44699 Published: 2026-10-30T13:08:02.000Z Updated: 2026-10-30T13:08:02.000Z Section: Technology > The September poll found 89% had limited or redesigned use cases because of token consumption; 45% had moved workloads in response to higher or more volatile costs. Cloudera released survey results showing 75% of respondents cited governance and compliance as barriers to scaling AI, while 89% limited or redesigned use cases amid rising token consumption. The September survey, commissioned by Cloudera and conducted by Wakefield Research, included 1,500 global IT leaders and decision-makers. The findings cover enterprise AI broadly, not AI-agent deployments specifically. More than half of respondents, 55%, said governance and compliance costs had grown faster than expected. Another 45% said they had moved AI workloads in response to higher or more volatile costs. Possible destinations included another cloud, a private environment or an on-premises platform, but the findings did not break down where workloads moved from or to. Data-location rules also influenced deployment choices. Eighty-four percent said data residency, sovereignty or jurisdiction-specific requirements affected where they deployed AI workloads. At Cloudera’s EVOLVE26 event in New York on Oct. 27, Chief Product Officer Leo Brunnick said, “The agent needs to know who you are and what your permissions are.” The survey did not measure agent deployments or how often permission requirements affect them. An April Cloudera survey of nearly 1,300 global IT leaders found nearly four in five respondents said limited data access across environments constrained AI and data initiatives. The surveys asked different questions, so their results do not provide a direct measure of change over time. The new findings add cost and governance context to [earlier coverage of companies’ plans for agent autonomy in production](). That survey found 27% of respondents already allowed some low-risk agents to deploy changes without human validation, while 20% were working toward that capability within 12 months. ## Links in this article - [earlier coverage of companies’ plans for agent autonomy in production](https://www.tokenpost.com/news/technology/29203)