# Run AI workloads on any cloud, store on Hugging Face: zero-egress storage with SkyPilot

Published: 2026-07-07 · Updated: 2026-07-07 · Category: Infrastructure

Source page: https://www.sipply.dev/story/run-ai-workloads-on-any-cloud-store-on-hugging-face-zero-86c17dea · 中文: https://www.sipply.dev/zh/story/run-ai-workloads-on-any-cloud-store-on-hugging-face-zero-86c17dea

## Summary

Hugging Face has partnered with SkyPilot to enhance the infrastructure for running AI workloads across various cloud platforms. This collaboration introduces a zero-egress storage solution, allowing users to store their datasets and models on Hugging Face while executing jobs on any cloud service without incurring additional egress fees. The integration simplifies the process of accessing and utilizing data, making it more efficient for developers and data scientists to manage their AI projects across different environments.

## Key takeaways

- Hugging Face and SkyPilot's partnership enables seamless AI workload management across clouds.
- The zero-egress storage solution eliminates additional costs associated with data transfer.
- Users can now easily access Hugging Face datasets and models directly within SkyPilot jobs.

## Why it matters

This development significantly reduces the cost and complexity of managing AI workloads across multiple cloud platforms, making it easier for organizations to leverage cloud resources effectively.

- **Product managers:** Product managers should note that this integration can enhance product offerings by providing customers with a cost-effective and efficient way to manage AI workloads, potentially increasing user adoption.
- **Developers:** For developers, this means less hassle with data management and reduced costs, allowing them to focus more on building and optimizing AI models rather than dealing with cloud storage limitations.
- **Students and job seekers:** Students and job-seekers can benefit from this advancement as it provides them with practical experience in using cutting-edge tools and technologies that facilitate AI development in a multi-cloud environment.

## Sources

- [Run AI workloads on any cloud, store on Hugging Face: zero-egress storage with SkyPilot](https://huggingface.co/blog/skypilot-hf-storage) (Hugging Face Blog, 2026-07-07)

## 中文

### 在任何云上运行AI工作负载，存储在Hugging Face：SkyPilot的零出口存储

Hugging Face与SkyPilot合作，增强了在各种云平台上运行AI工作负载的基础设施。此次合作推出了零出口存储解决方案，允许用户在Hugging Face上存储数据集和模型，同时在任何云服务上执行任务而无需支付额外的出口费用。该集成简化了数据访问和利用的过程，使开发人员和数据科学家在不同环境中管理AI项目更加高效。

**要点：**

- Hugging Face与SkyPilot的合作实现了跨云的无缝AI工作负载管理。
- 零出口存储解决方案消除了与数据传输相关的额外成本。
- 用户现在可以在SkyPilot任务中直接访问Hugging Face的数据集和模型。

**为什么重要：** 这一发展显著降低了在多个云平台上管理AI工作负载的成本和复杂性，使组织能够更有效地利用云资源。

---
Sipply: Sip AI News Simply. https://www.sipply.dev