HPE is pressing enterprises to rethink how they pay for AI as workloads move from pilots to production, arguing that consumption pricing stops making sense once demand becomes steady and business-critical. In a sponsored piece published by MIT Technology Review, HPE's Cheri Williams frames the choice as a workload-by-workload decision between renting tokens and owning capacity.
The pitch rests on a shift Williams says is already underway. She cites Deloitte's 2026 State of AI in the Enterprise, which found worker access to AI rose 5% in 2025, and projects that the share of companies with at least 40% of their AI projects in production will double within six months.
“When demand becomes steady and business-critical, a consumption-only approach can turn AI spending into a variable monthly line item that is difficult to forecast as usage, workloads, and model requirements change.”— Cheri Williams, HPE
The argument: assistants, retrieval-and-knowledge systems, and agentic applications create recurring demand across models, data, and tools. Once that demand is predictable and large enough to keep infrastructure productive, Williams says leaders should ask whether it still makes economic sense to buy AI one request at a time.
Key facts
- 01Deloitte's 2026 State of AI in the Enterprise reports worker access to AI rose 5% in 2025.
- 02The share of companies with at least 40% of AI projects in production is expected to double within six months.
- 03HPE's Cheri Williams urges leaders to model AI demand over the next 12 to 18 months before committing capital.
- 04The argument frames owned AI infrastructure as a productive asset only when utilization is sustained.
She stops short of declaring ownership the cheaper option by default. Every organization has a crossover point that depends on model choice, input-output token balance, performance requirements, system design, energy costs, and the operating model needed to support the workload. Retrieval-heavy systems and agentic workflows carry very different cost profiles from a simple assistant.
“Ownership is not automatically the lower-cost answer. It only makes sense when an enterprise can keep capacity productive.”— Cheri Williams, HPE
Williams outlines three questions for leaders weighing the shift: whether demand is steady enough to justify dedicated capacity, at what utilization level ownership becomes economical, and whether the organization can keep that capacity productive through adoption, governance, and continued use-case expansion. The framing sets ownership as a discipline rather than a default.
The piece was produced by HPE, whose infrastructure business stands to benefit if enterprises accept the premise. Williams closes by arguing that the organizations that create the most value will know when recurring demand calls for a different economic model — the point at which, she writes, AI stops being an expense and becomes an asset.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
