Operated by its subsidiary Shenzhen Junyu Intelligent Computing Technology Co., Ltd. — a full-stack AI service provider whose core members come from IBM, Tongji University, Japan and other leading institutions worldwide. It provides on-demand leasing of GPU compute resources and turnkey cluster delivery for workloads such as large-model training, inference and fine-tuning; a professional team handles networking, hardware-software co-tuning and factory acceptance, with compute capacity scaling elastically.
Compared with a pure resale model, this path shortens the delivery chain, reduces intermediate mark-ups and supports hardware-software co-tuning for each customer's scenario.
Compute resources are supplied elastically according to workload, avoiding one-off heavy-asset investment.
Customised configuration and cluster delivery for workloads such as large-model training, inference and fine-tuning.
Network integration and factory acceptance testing before delivery, with continuous performance tuning afterwards.
| Workloads served | Large-model training / inference / fine-tuning |
|---|---|
| Delivery format | Turnkey cluster delivery |
| Supply model | On-demand leasing · elastic scaling |
| High-speed interconnect | InfiniBand / RoCE |