PRODUCTS
Products
From 1P entry-level computing to enterprise heterogeneous clusters
WeCalc delivers unified-architecture, tiered solutions for every scale of on-premise AI
WeCalc-B Basic
1P entry-level computing — launch your local AI pilot fast
WeCalc-B is built on a "minimum viable" philosophy, integrating compute, storage, and management into one device so enterprises and universities can establish a secure, local, and scalable AI computing base within 48–72 hours.
Ideal Customer Types
- Enterprises wanting to pilot first
- Budget-conscious teams
- University teaching & training
Recommended Industries
Lease from ¥2,000/month
Core Value
- A single unit can launch full computing services — ideal for pilot validation and lightweight deployment.
- All data stays on your own devices, meeting data-sovereignty requirements for education, government, and R&D.
Delivery Model
- Ships with integrated hardware and software — ready to deploy on arrival.
- Supports one-click startup and remote O&M, reducing on-site implementation complexity.
WeCalc-P Professional
Mid-scale cluster for training and inference
WeCalc-P leverages multi-CPU, multi-GPU node clusters with EBOF all-flash storage architecture to deliver higher throughput, lower latency, and stronger scalability for training, inference, and edge analytics workloads.
Ideal Customer Types
- Customers ready for production
- Teams with defined use cases
- Mid-scale training & inference projects
Recommended Industries
Most popular production-grade solution
Core Value
- Combines disaggregated storage-compute with all-flash storage to significantly boost data loading efficiency and throughput.
- Bridges the gap from pilot to production with balanced performance, cost, and scalability.
Delivery Model
- Pre-configured per scenario to reduce on-site integration time.
- Provides compute, storage, and networking as an integrated turnkey delivery with O&M support.
WeCalc-E Enterprise
Thousand-GPU heterogeneous cluster for HPC and large-scale training
WeCalc-E targets high-density, high-throughput, and high-reliability workloads. Supporting multi-node heterogeneous clusters and PB-level distributed storage pools, it is the flagship solution for enterprises and research institutions building on-premise computing infrastructure.
Ideal Customer Types
- Regional intelligent computing centers
- Large-scale training platforms
- Research computing platforms
Recommended Industries
Custom delivery based on business scale
Core Value
- Handles ultra-large-scale training and HPC tasks, meeting demands for high concurrency, massive data, and complex models.
- High-speed interconnect and distributed storage ensure data throughput and cluster stability for large workloads.
Delivery Model
- Complete solution design for compute, networking, and storage based on business objectives.
- Full-lifecycle delivery from site assessment to cluster go-live.
FEATURES
Core Features
From plug-and-play to modular expansion — WeCalc reduces computing center delivery complexity to a range that fits real enterprise deployments
Plug and Play
Ships with fully integrated hardware and software — no specialized setup required.
Single-Unit Operation
One device delivers a complete computing service, ideal for quick pilots.
One-Click Startup
Power on and launch computing services with one click, lowering deployment barriers.
Turnkey Delivery
End-to-end turnkey service from deployment to operation, completed in 48–72 hours.
Modular Expansion
Supports hot-swap expansion, scaling smoothly from a single unit to a cluster.
Multi-Hardware Support
Flexibly accommodates CPUs, GPUs, SSDs, and more.
Domestic HW Compatible
Huawei Ascend & Kunpeng certified, compatible with 90%+ of domestic GPUs.
COMPARISON
Product Specification Comparison
| Specification | WeCalc-B Basic | WeCalc-P Professional Recommended | WeCalc-E Enterprise |
|---|---|---|---|
| Compute | 1×CPU + Optional GPU | Multi-CPU + Multi-GPU Cluster | Thousand-GPU Heterogeneous Cluster |
| Storage | 4TB NVMe SSD | 16×3.84TB EBOF | PB-level Distributed Storage |
| Computing Power | ≤1 PFLOPS | ≤12 PFLOPS | ≥50 PFLOPS |
| Network | 25G/100G Ethernet | 100G RDMA | 200G/400G Interconnect |
| Deployment | 48–72 hours | 48–72 hours | Custom schedule |
| Scalability | Single unit, linear scaling | Up to 100 nodes | 10,000+ node evolution |
| Domestic HW Compatible | ✓ | ✓ | ✓ |
| On-Premise Data | ✓ | ✓ | ✓ |
| Reference Price | ¥98K / ¥2,000/mo | ¥2–5 million | ¥5 million+ |
Financing Lease
Starting from just ¥2,000/month for 1P of computing power
Equivalent to approximately ¥40,000 in ChatGPT token credits — ideal for enterprises and institutions looking to adopt local AI capabilities with minimal upfront investment.
Find the Right WeCalc Product for You
Our expert team will recommend the best solution based on your actual needs
Free Consultation