Unverified50% confidenceFactExact time
For usage-based AI products, NRR is easily dragged down by customers' usage-optimization behavior, because customers learning to use the product in a more token-efficient way actually leads to revenue contraction.
1
Sources
50%
Confidence
Long-term
Relevance
10/22/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedFor AI companies, inference costs go directly into cost of goods sold, which depresses gross margin and in turn depresses LTV.74% similarUnverifiedIn a provider context, 'model efficiency gains' usually means lower inference cost per token, which users only perceive when the efficiency gains translate into higher quotas or lower prices.72% similarUnverifiedThe industry generally believes that the low switching cost of AI applications leads to extremely rapid user churn, and the retention curve of purely general-purpose Q&A products often drops sharply in the following month.71% similarUnverifiedFor AI companies, there is debate over whether revenue generated by usage-based billing counts as ARR, because it fluctuates with usage and is less predictable than seat subscriptions.71% similarUnverifiedAI lowers the barrier to penetration testing, and also lowers the barrier to abuse.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1086243API
curl https://kongchang.com/api/v1/knowledge/claims/1086243MCP
get_claim(id=1086243)