DeepSeek's DSec paper: 3M sandboxes/day, 380K concurrent, powering V3.2-V4.1 RL
jiqizhixin · x · 2026-10-02
DeepSeek (with Liang Wenfeng among 100+ authors) published DSec (DeepSeek Elastic Compute), a 31-page systems paper on its production RL sandbox platform: 160 CPU nodes, 30,000 cores, 250 TB DRAM, PB-scale images and environment layers, 3 million sandbox instances per day, 380K+ concurrent at peak, created at over 5,000/second — running every RL training and evaluation sandbox from V3.2 to V4.1.
The core question: as agent training scales to hundreds of thousands of concurrent agents, where do they run? Each needs an independent sandbox holding code, dependencies, test tools and services, with file/process state persisting across multi-round interactions — traditional token-generating inference can't carry that state.
More from Infra
- Micron Q4 FY26: bigger bit base and more HBM support a longer memory cycle — BenBajarin · 2026-10-02
- AWS revamps onboarding: no credit card, $200 free credits, hard spend caps, one-prompt agent setup — danbri · 2026-10-02
- Lambda wins ClusterMAX Sliver as SemiAnalysis finds reliability much improved — TheZachMueller · 2026-10-02
- Micron turns green as AI compute selloff appears to cool — firstadopter · 2026-10-02
- Edge0 open-sources app layer: 35B model on-device with just 1–2.5GB memory — FinanceYF5 · 2026-10-02
- SpaceX Transporter-18 launches 130 payloads including Google's orbital AI chips — XFreeze · 2026-10-02