10 questions that will make IR wish you’d shut up about AI

1. AI companies talk endlessly about revenue growth — is AI cost rising faster than AI revenue?
2. If an AI system rehydrates context on every turn and unit cost rises with usage, how is that scaling rather than a cost curve pointing the wrong way?
3. If token pricing is only a proxy, why are the real drivers of AI cost — compute, memory, orchestration — not disclosed directly?
4. How can “AI margins” be evaluated without publishing the marginal cost of one additional query?
5. If expanding context windows increase latency, parsing overhead, and compute load, on what basis is that being described as efficiency?
6. When earnings calls emphasize “usage” but omit unit economics, what economic model are investors supposed to be evaluating?
7. If most AI case studies assume unlimited budget, what happens to the business case once real enterprise budget constraints are applied?
8. If an AI system always runs inference, where is the cost governance that distinguishes intelligence from an expensive reflex loop?
9. If AI cost control cannot be fixed downstream with billing, why is execution not governed before inference, when costs are actually incurred
10. If AI is as transformative as claimed, why are cost breakdowns avoided — and when can investors expect revenue per seat?

– Published Thursday, January 29, 2026