# Thousand-GPU Large-Scale Training and Optimization Recipe for AI-Native Cloud Embodied Intelligence Infrastructure

arXiv: https://arxiv.org/abs/2603.11101 · HTML: https://mlsanity.com/paper/2603.11101/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 31 references that are in the map.
- First posted on arXiv 11 Mar 2026.

## Builds on (most cited first, 25 of 25)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [π0: A Vision-Language-Action Flow Model for General Robot Control](https://mlsanity.com/paper/2410.24164/) (2024)
- [High-Dimensional Continuous Control Using Generalized Advantage Estimation](https://mlsanity.com/paper/1506.02438/) (2015)
- [Mastering Diverse Domains through World Models](https://mlsanity.com/paper/2301.04104/) (2023)
- [GR00T N1: An Open Foundation Model for Generalist Humanoid Robots](https://mlsanity.com/paper/2503.14734/) (2025)
- [Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm](https://mlsanity.com/paper/1712.01815/) (2017)
- [Mobile ALOHA: Learning Bimanual Mobile Manipulation with Low-Cost Whole-Body Teleoperation](https://mlsanity.com/paper/2401.02117/) (2024)
- [SimpleVLA-RL: Scaling VLA Training via Reinforcement Learning](https://mlsanity.com/paper/2509.09674/) (2025)
- [AI Agents Under Threat: A Survey of Key Security Challenges and Future Pathways](https://mlsanity.com/paper/2406.02630/) (2024)
- [DeepSpeed-Chat: Easy, Fast and Affordable RLHF Training of ChatGPT-like Models at All Scales](https://mlsanity.com/paper/2308.01320/) (2023)
- [MuJoCo Playground](https://mlsanity.com/paper/2502.08844/) (2025)
- [VideoVLA: Video Generators Can Be Generalizable Robot Manipulators](https://mlsanity.com/paper/2512.06963/) (2025)
- [RLinf: Flexible and Efficient Large-scale Reinforcement Learning via Macro-to-Micro Flow Transformation](https://mlsanity.com/paper/2509.15965/) (2025)
- [ChatVLA-2: Vision-Language-Action Model with Open-World Embodied Reasoning from Pretrained Knowledge](https://mlsanity.com/paper/2505.21906/) (2025)
- [Dexbotic: Open-Source Vision-Language-Action Toolbox](https://mlsanity.com/paper/2510.23511/) (2025)
- [Prepacking: A Simple Method for Fast Prefilling and Increased Throughput in Large Language Models](https://mlsanity.com/paper/2404.09529/) (2024)
- [Hummer: Towards Limited Competitive Preference Dataset](https://mlsanity.com/paper/2405.11647/) (2024)
- [Ring-lite: Scalable Reasoning via C3PO-Stabilized Reinforcement Learning for LLMs](https://mlsanity.com/paper/2506.14731/) (2025)
- [RL-VLA3: A Flexible and Asynchronous Reinforcement Learning Framework for VLA Training](https://mlsanity.com/paper/2602.05765/) (2026)
- [SHARP: Synthesizing High-quality Aligned Reasoning Problems for Large Reasoning Models Reinforcement Learning](https://mlsanity.com/paper/2505.14147/) (2025)
- [Robot Learning: A Tutorial](https://mlsanity.com/paper/2510.12403/) (2025)
- [How Social is It? A Benchmark for LLMs' Capabilities in Multi-user Multi-turn Social Agent Tasks](https://mlsanity.com/paper/2505.04628/) (2025)
- [Learning Truthful Mechanisms without Discretization](https://mlsanity.com/paper/2506.22911/) (2025)

## Built on by (most cited first, 3 of 3)

- [Pre-VLA: Preemptive Runtime Verification for Reliable Vision-Language-Action and World-Model Rollouts](https://mlsanity.com/paper/2605.22446/) (2026)
- [Sword: Style-Robust World Models as Simulators via Dynamic Latent Bootstrapping for VLA Policy Post-Training](https://mlsanity.com/paper/2605.07288/) (2026)
- [D-VLA: A High-Concurrency Distributed Asynchronous Reinforcement Learning Framework for Vision-Language-Action Models](https://mlsanity.com/paper/2605.13276/) (2026)

## Abstract (arXiv)

Embodied intelligence is a key step towards Artificial General Intelligence (AGI), yet its development faces multiple challenges including data, frameworks, infrastructure, and evaluation systems. To address these issues, we have, for the first time in the industry, launched a cloud-based, thousand-GPU distributed training platform for embodied intelligence, built upon the widely adopted LeRobot framework, and have systematically overcome bottlenecks across the entire pipeline. At the data layer, we have restructured the data pipeline to optimize the flow of embodied training data. In terms of training, for the GR00T-N1.5 model, utilizing thousand-GPU clusters and data at the scale of hundreds of millions, the single-round training time has been reduced from 15 hours to just 22 minutes, achieving a 40-fold speedup. At the model layer, by combining variable-length FlashAttention and Data Packing, we have moved from sample redundancy to sequence integration, resulting in a 188% speed increase; π-0.5 attention optimization has accelerated training by 165%; and FP8 quantization has delivered a 140% speedup. On the infrastructure side, relying on high-performance storage, a 3.2T RDMA network, and a Ray-driven elastic AI data lake, we have achieved deep synergy among data, storage, communication, and computation. We have also built an end-to-end evaluation system, creating a closed loop from training to simulation to assessment. This framework has already been fully validated on thousand-GPU clusters, laying a crucial technical foundation for the development and application of next-generation autonomous intelligent robots, and is expected to accelerate the arrival of the era of human-machine integration.