# Who Pays for the KV Cache? Attributing Shared AI Inference Spend Across Kubernetes and LLM Provider Bills

arXiv: https://arxiv.org/abs/2609.24991 · HTML: https://mlsanity.com/paper/2609.24991/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Serving systems and request scheduling, one of 12 directions in Efficient LLM inference and serving.
- Builds on 14 of its 24 references that are in the map.
- First posted on arXiv 21 Sep 2026.

## Placed in

- Efficient LLM inference and serving / [Serving systems and request scheduling](https://mlsanity.com/inference/serving-systems-and-request-scheduling/)

## Builds on (most cited first, 14 of 14)

- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://mlsanity.com/paper/1909.08053/) (2019)
- [SGLang: Efficient Execution of Structured Language Model Programs](https://mlsanity.com/paper/2312.07104/) (2023)
- [DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving](https://mlsanity.com/paper/2401.09670/) (2024)
- [Efficiently Scaling Transformer Inference](https://mlsanity.com/paper/2211.05102/) (2022)
- [Taming Throughput-Latency Tradeoff in LLM Inference with Sarathi-Serve](https://mlsanity.com/paper/2403.02310/) (2024)
- [Carbon accounting in the Cloud: a methodology for allocating emissions across data center users](https://mlsanity.com/paper/2406.09645/) (2024)
- [ABACUS: A FinOps Service for Cloud Cost Optimization](https://mlsanity.com/paper/2501.14753/) (2024)
- [Preserving Admission Responsibility in Multi-Tenant Large Language Model Prefix Caches](https://mlsanity.com/paper/2608.01657/) (2026)
- [Characterization of Request and Token Energy Costs for LLM Inference Workloads on GPU Platforms](https://mlsanity.com/paper/2608.28044/) (2026)
- [Cost-Governed RAG: Unified Per-Tenant Cost Attribution Across Retrieval and Generation in Multi-Tenant LLM Systems](https://mlsanity.com/paper/2607.12188/) (2026)
- [Request-Level Energy Attribution for Batched LLM Serving](https://mlsanity.com/paper/2608.00026/) (2026)
- [LLMVisor: A Real-Time Latency Attribution Model for Multi-Tenant LLM Serving](https://mlsanity.com/paper/2608.08382/) (2026)

## Abstract (arXiv)

Organizations pay for AI through disconnected ledgers: Kubernetes allocations for self-hosted inference, gateway logs, and per-token bills from API providers. We present unalloc, an open-source tool that joins OpenCost, LiteLLM, OpenAI and Anthropic cost data into one exact ledger and reports the share of spend with no owner, and use it to study where attribution breaks at the seams between these systems. Five case studies run inference for real or simulate it: a vLLM-style serving simulator with paged KV memory and prefix caching; a PyTorch transformer serving a multi-tenant trace with a real KV cache; tensor- and pipeline-parallel inference on torch.distributed; the unmodified CLI against mock provider APIs; and four downstream use cases. At the seams, in a constructed multi-pod deployment scenario -- one month of synthetic OpenCost allocations, not observed billing data -- owner labels set only on LeaderWorkerSet leader pods leave 66% of that deployment's GPU bill unowned, and the natural fallback key assigns 61% of it to a Helm chart name while the headline unallocated share falls to 4%; enabling every source double counts all gateway spend; and reading one page of a billing API reports a quarter of spend. Inside a shared inference server the metering rule decides who pays: on an NVIDIA H100 running vLLM, a token meter assigns a retrieval-heavy tenant 12-14 percentage points more of the bill than an equal time-share meter at every load tested, while GPU utilization reads 97-99% across configured loads of 2 to 16 requests per second (3.7 to 26.9 completed requests per second; the configured rate counts session-initial arrivals only) and power draw tracks load. Neither meter is a ground truth; we position these results against recent Shapley-based energy attribution. Code, raw data, captured evidence, figures and the paper regenerate from the repository.