Most finance and technology leaders are still measuring AI the same way they measure a utility bill: by counting tokens. It's an easy number to track. It's also the wrong number.
On October 29, 2026, Sreenivasan “Raj” Rajagopal, GM and head of ValueOps AI and infrastructure, will be joined by Chris Wolf, global head of AI and advanced services, VMware Cloud Foundation Division, for SD Times' Economics of AI Supercast. At this event, we're making the case for a different way to think about AI economics—one built around the job AI is actually doing, not the tokens it burns getting there.
Here’s a preview of what we'll cover:
The token price isn't the true cost. What hidden infrastructure and operational expenses have an impact on total enterprise AI costs? Add observability, backup, and availability to the equation, and the number on your invoice is only the visible tip of what AI actually costs you.
Cost-per-token is the wrong unit. Cost-per-task is the right one. Why is cost-per-task a better metric than token costs for measuring the ROI of enterprise AI? You don't measure a contractor by hours logged—you measure the job they were hired to do. The same logic applies to AI, and measuring it that way makes some uncomfortable questions a lot easier to ask.
Intelligence inside your organization isn't exclusively human anymore. That shift to artificial intelligence has to be managed deliberately, with the same fiscal discipline you'd apply to any other investment.
You don't have to transform everything at once. How should CFOs and technology leaders budget for long-term AI transformation? AI transformation is a multi-year journey, not a single leap. The good news is that there's a real, staged way to start that gets your board a number it can actually evaluate.
If you're a CFO, CIO, or FinOps leader trying to get ahead of AI spending before it gets ahead of you, this is 45 minutes worth carving out.
Register for the Economics of AI Supercast
To continue the discussion, please contact us for a demo or conversation.
A: Measuring AI spending solely by token count is misleading because token pricing represents only the visible surface of an invoice. These figures omit such essential expenses as investments needed to manage observability, backup, and availability. Similar to evaluating a contractor by the completed project rather than hours logged, AI should be assessed by the job it delivers rather than raw tokens consumed.
A: The recommended metric for evaluating enterprise AI economics is cost-per-task rather than cost-per-token. Shifting the focus to the specific task AI accomplishes provides an accurate picture of return on investment and makes critical spending evaluations easier to address.
A: Enterprise leadership should approach AI transformation as a staged, multi-year journey rather than a single, all-at-once overhaul. Phasing adoption into distinct stages establishes realistic milestones and equips corporate boards with verifiable numbers they can properly evaluate.
A: The Economics of AI Supercast takes place on October 29, 2026. The event features Sreenivasan “Raj” Rajagopal, GM and head of ValueOps AI and infrastructure and Chris Wolf, global head of AI and advanced services, VMware Cloud Foundation Division. The 45-minute event is tailored for CFOs, CIOs, and FinOps leaders looking to manage enterprise AI expenditures proactively.