# Remoe: Towards Efficient and Low-Cost MoE Inference in Serverless Computing

arXiv: https://arxiv.org/abs/2512.18674 · HTML: https://mlsanity.com/paper/2512.18674/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Efficient MoE training and inference systems, one of 12 directions in Mixture-of-experts.
- Builds on 17 of its 38 references that are in the map.
- First posted on arXiv 21 Dec 2025.

## Placed in

- Mixture-of-experts / [Efficient MoE training and inference systems](https://mlsanity.com/moe/efficient-moe-training-and-inference-systems/)

## Builds on (most cited first, 17 of 17)

- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models](https://mlsanity.com/paper/2401.06066/) (2024)
- [DeepSeek-V3 Technical Report](https://mlsanity.com/paper/2412.19437/) (2024)
- [Phi-4 Technical Report](https://mlsanity.com/paper/2412.08905/) (2024)
- [LMSYS-Chat-1M: A Large-Scale Real-World LLM Conversation Dataset](https://mlsanity.com/paper/2309.11998/) (2023)
- [MoE-LLaVA: Mixture of Experts for Large Vision-Language Models](https://mlsanity.com/paper/2401.15947/) (2024)
- [HOBBIT: A Mixed Precision Expert Offloading System for Fast MoE Inference](https://mlsanity.com/paper/2411.01433/) (2024)
- [ProMoE: Fast MoE-based LLM Serving using Proactive Caching](https://mlsanity.com/paper/2410.22134/) (2024)
- [Taming Latency-Memory Trade-Off in MoE-Based LLM Serving via Fine-Grained Expert Offloading](https://mlsanity.com/paper/2502.05370/) (2025)
- [ExpertFlow: Efficient Mixture-of-Experts Inference via Predictive Expert Caching and Token Scheduling](https://mlsanity.com/paper/2410.17954/) (2024)
- [Lynx: Enabling Efficient MoE Inference through Dynamic Batch-Aware Expert Selection](https://mlsanity.com/paper/2411.08982/) (2024)
- [HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference](https://mlsanity.com/paper/2504.05897/) (2025)
- [HydraServe: Minimizing Cold Start Latency for Serverless LLM Serving in Public Clouds](https://mlsanity.com/paper/2502.15524/) (2025)
- [ENOVA: Autoscaling towards Cost-effective and Stable Serverless LLM Serving](https://mlsanity.com/paper/2407.09486/) (2024)
- [MoE-GPS: Guidlines for Prediction Strategy for Dynamic Expert Duplication in MoE Load Balancing](https://mlsanity.com/paper/2506.07366/) (2025)
- [MOPAR: A Model Partitioning Framework for Deep Learning Inference Services on Serverless Platforms](https://mlsanity.com/paper/2404.02445/) (2024)
- [Towards Resource-Efficient Serverless LLM Inference with SLINFER](https://mlsanity.com/paper/2507.00507/) (2025)

## Built on by (most cited first, 2 of 2)

- [MoEless: Efficient MoE LLM Serving via Serverless Computing](https://mlsanity.com/paper/2603.06350/) (2026)
- [OLED-MoE: Accelerating MoE-Based dLLM Inference via Inter-Iteration Locality-Aware Expert Offloading](https://mlsanity.com/paper/2609.33385/) (2026)

## Abstract (arXiv)

Mixture-of-Experts (MoE) has become a dominant architecture in large language models (LLMs) due to its ability to scale model capacity via sparse expert activation. Meanwhile, serverless computing, with its elasticity and pay-per-use billing, is well-suited for deploying MoEs with bursty workloads. However, the large number of experts in MoE models incurs high inference costs due to memory-intensive parameter caching. These costs are difficult to mitigate via simple model partitioning due to input-dependent expert activation. To address these issues, we propose Remoe, a heterogeneous MoE inference system tailored for serverless computing. Remoe assigns non-expert modules to GPUs and expert modules to CPUs, and further offloads infrequently activated experts to separate serverless functions to reduce memory overhead and enable parallel execution. We incorporate three key techniques: (1) a Similar Prompts Searching (SPS) algorithm to predict expert activation patterns based on semantic similarity of inputs; (2) a Main Model Pre-allocation (MMP) algorithm to ensure service-level objectives (SLOs) via worst-case memory estimation; and (3) a joint memory and replica optimization framework leveraging Lagrangian duality and the Longest Processing Time (LPT) algorithm. We implement Remoe on Kubernetes and evaluate it across multiple LLM benchmarks. Experimental results show that Remoe reduces inference cost by up to 57% and cold start latency by 47% compared to state-of-the-art baselines.