Enterprise AI scalability is facing significant budget constraints as companies exhaust cloud token allocations much faster than anticipated. Dell Technologies and Intel report that businesses are now reevaluating their deployment strategies to manage these rising costs.
Helen O’Sullivan, AI business development manager at Dell, noted that many customers in the Bay Area are depleting months-long token budgets within just a few weeks. To mitigate this, firms are increasingly looking toward on-premises infrastructure and open-weight models.
The rise of agentic workloads is shifting hardware requirements, with a growing emphasis on CPU utilization alongside traditional GPU demand. Intel’s Tim Wood highlighted that many enterprises already possess existing CPU resources in their data centers capable of supporting next-generation applications.
Dell and Intel are currently advising clients to assess their entire technology stack before production. This consultative approach involves evaluating the balance between capital expenditures and operating costs, as well as testing models in sandbox environments to determine actual workload needs.
