Unverified50% confidenceFactExact time
The inference cost of large models grows linearly or even super-linearly with business scale, and cannot serve more users simply by copying code.
1
Sources
50%
Confidence
Long-term
Relevance
9/29/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe relationship between inference cost and model scale is not linear; when the parameter count doubles, inference latency and computational consumption often increase by far more than double.75% similarUnverifiedMany AI vendors' premium subscriptions operate at a loss, because each large language model inference requires matrix computations across billions to hundreds of billions of parameters, with costs growing linearly or even super-linearly with usage.71% similarUnverifiedReasoning models may consume several times or even more than ten times the output tokens, directly driving up API call costs.71% similarUnverified推理成本（Inference Cost）是大模型持续失血的主要原因，随用户规模呈线性甚至超线性增长70% similarUnverified在反向模式下，当模型参数数量高达数百万时，计算梯度的成本仅与一次前向计算相当69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/958338API
curl https://kongchang.com/api/v1/knowledge/claims/958338MCP
get_claim(id=958338)