Dell CFO Says AI Budgets Must Shift From Token Costs To Speed And Productivity
Dell’s David Kennedy says finance chiefs are treating AI as a production budget line, prioritizing speed over simple cost control.
Key points
- Dell Technologies reports 6,500 AI factories across its customer base, signaling mass adoption.
- Financial services firms are leading the shift from AI experimentation to scaled production.
- CFOs are creating dedicated budget lines for token spend to manage variable costs.
Dell Technologies Inc. chief financial officer David Kennedy stated that enterprise AI has moved past the experimentation phase into full production, requiring financial discipline. According to a report by SiliconANGLE, finance leaders are now treating artificial intelligence as a standard budget line item rather than a science project, focusing on speed and productivity gains to justify expenditure.
Kennedy emphasized that while token consumption costs require management, they will not be the primary value driver for enterprises. He told theCUBE Research’s Dave Vellante that IT and finance teams should establish guardrails around spending, but the real competitive advantage comes from the speed and efficiency AI delivers to the organization. This perspective was shared during an exclusive broadcast at the Dell Technologies New York Symposium.
Scaling AI in Financial Services
Financial services companies, particularly those involved in high-frequency trading, are among the first industries to prove that AI can deliver results at scale. Kennedy noted that pharmaceutical companies are likely the next sector to follow this trajectory. Dell Technologies currently counts 6,500 AI factories within its customer base, indicating widespread deployment. In May, the company expanded its AI Factory partnership with Nvidia to help more customers bridge the gap between initial experiments and measurable business outcomes.
Budgeting and Model Placement
The shift toward production involves basic financial discipline, with most chief financial officers now creating dedicated line items for token spend. Kennedy explained that tracking these costs like any other operational expense is crucial for scaling effectively. He suggested that efficiency often depends on model placement, noting that repetitive finance workflows can run on open-weight models on-premises. This strategy helps organizations turn variable token costs into fixed expenses, providing greater budgetary predictability.
What should you do?
Organizations should audit their current AI spending to distinguish between experimental costs and production-level expenses. Finance and IT teams need to collaborate on establishing clear guardrails for token consumption. Companies should evaluate whether certain workflows can be moved to on-premises open-weight models to reduce variable costs. Leaders should focus on measuring the speed and productivity gains AI delivers rather than just the raw cost of tokens.
What to do and how to stay safe: Dell Technologies
- Create a dedicated budget line item specifically for AI token consumption to track spending accurately.
- Collaborate with IT departments to establish guardrails that manage token usage without stifling innovation.
- Evaluate repetitive workflows to determine if they can run on open-weight models on-premises to fix costs.
- Measure AI initiatives based on speed and productivity improvements rather than just initial implementation costs.
Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
How many AI factories does Dell Technologies currently have?
Dell Technologies counts 6,500 AI factories across its customer base, according to CFO David Kennedy.
Which industries are leading the adoption of AI at scale?
Financial services firms, especially in high-frequency trading, are early adopters, with pharmaceutical companies likely next.
How should CFOs manage AI token costs?
CFOs should create dedicated line items for token spend and consider running repetitive workflows on open-weight models on-premises to turn variable costs into fixed ones.
