# DeepSeek details DSec sandbox infrastructure for agent training — retained record

> Retained china-news record from china-news:technode.

Canonical: https://highsignal.app/data/records/3c1915c576c43908

[← sources](https://highsignal.app/data)


source: [china-news](https://highsignal.app/data/china-news) (china-news:technode)  published 2026-09-23 09:14 ingested 2026-09-24 02:50

[original source ↗](https://technode.com/2026/09/23/deepseek-dsec-agent-training-sandbox-infrastructure/)

## retained text



```
A paper posted to arXiv details DeepSeek Elastic Compute, or DSec, a sandbox platform designed for large-scale agent training. DeepSeek founder Liang Wenfeng is listed among the paper’s 130-plus authors.
DSec unifies function-call, container, microVM and full-VM sandboxes, while coordinating their lifecycle with reinforcement learning workloads. The paper says a production-scale unit spans around 160 nodes, supports about 3 million sandboxes per day and more than 380,000 concurrent sandboxes, and can create more than 5,000 sandboxes per second. [arXiv]
```



document fetched 2026-09-26 02:47 · canonical https://technode.com/2026/09/23/deepseek-dsec-agent-training-sandbox-infrastructure
