Cloud Token Costs Force Enterprises to Rethink Agentic AI Hardware

Enterprise IT departments are hitting strict budget constraints as autonomous AI agents consume cloud tokens significantly faster than anticipated. The resulting cost pressure is driving organizations to reconsider hybrid strategies, pairing on-premises infrastructure and open-weight models with cloud frontier systems, according to reporting by SiliconANGLE (https://siliconangle.com/2026/10/12/dell-intel-weigh-costs-enterprise-ai-scalability-dellaileadershipsymposium/) during Dell's AI Leadership Symposium on theCUBE.

Helen O’Sullivan, AI business development manager and solutions specialist at Dell Technologies Inc., noted that enterprise customers in Northern California are burning through quarterly resource allocations in weeks. "A lot of our customers I support in the Bay Area, they’re dealing with this real immediate struggle of running through their token budget in a matter of weeks that they had allotted for months," O’Sullivan said. "You still use those frontier models, but you use the open source and infrastructure on-prem to at least eliminate some of the token stress and token maxing going on right now."

This shift is altering how enterprises evaluate hardware demands. While procurement teams remain focused on securing graphics processing units, agentic workloads also place substantial operational demands across central processing units for orchestration, inference, and edge processing.

Tim Wood, sales director for the Northwest region at Intel Corp., pointed out that organizations often have untapped compute capacity already installed. "A lot of enterprises already have CPUs in their data center. There may be existing resources and headroom in the infrastructure they already have that’s capable of doing some of these next-generation apps and workloads," Wood said, advising IT teams to evaluate open-weight models in sandbox environments before committing to new capital deployments.

Dell and Intel representatives both highlighted that enterprise AI scalability requires auditing end-to-end architectures rather than focusing solely on accelerator chips. "A lot of [customers] just see the [graphics processing unit] and [ask], ‘How fast can I get the GPU?’ or Vera Rubin or what have you," O’Sullivan said. "But it’s about the data."