MicroAGI Unveils Petabyte-Scale Robot Data Delivery Platform
German embodied AI startup MicroAGI publicly described a two-hop platform for delivering large multimodal robot-training datasets to S3-compatible object storage. The first hop stages source data in Hugging Face storage; the second uses Cloudflare Workers and Queues to transfer it to a customer's storage. MicroAGI said the process does not share customer credentials and can deliver petabyte-scale datasets. The company also reported throughput, cost, integrity-checking and dataset-browsing capabilities, but these figures have not been independently verified in the supplied report.
The platform is designed to move large robot-training datasets into customers' own cloud environments while avoiding shared customer credentials, addressing storage fragmentation and data-delivery requirements in robot learning.
10月28日消息，近日，德国具身 AI 初创公司 MicroAGI 对外公开自研海量机器人训练数据交付平台，采用两跳中转架构，可实现每日拍字节（Petabyte）级多模态机器人数据分发，能够将百万小时级视频、深度、姿态数据集投递至任意兼容 S3 协议的对象存储，解决机器人学习场景下数据集体量庞大、客户训练云环境分散的现实痛点。
机器人训练数据集动辄数百万文件、数 PB 规模，不同团队分别使用 AWS、GCP、Azure、Cloudflare R2、Hugging Face 等不同存储服务。平台确立三条核心设计原则：兼容几乎所有 S3 类对象存储后端；多租户任务之间互不抢占带宽资源；全程不共享客户密钥，访问权限严格绑定单次交付任务。
整套链路分为第一跳：源存储到 Hugging Face 中转存储、第二跳：中转存储推送到客户目标业务存储。
第一跳将原始数据统一暂存至 Hugging Face Bucket 作为中立中转层。相同数据集可被多次复用，源数据仅读取一次即可反复分发给不同客户、不同地域。底层依托 Xet 内容寻址分块存储，内容重复的文件直接复用已有分块，大幅降低网络传输量。
#MicroAGI#Hugging Face#Hugging Face Xet#Cloudflare Workers#Cloudflare Queues#AWS#Google Cloud Platform#Microsoft Azure#Cloudflare R2#Viewer