GPUStack Developer Shuche Software Raises ¥50M Seed Led by Shuihua Weilai | VeilStrat

GPUStack Parent Shuche Software Closes ¥50 Million Seed Round

Shenzhen-based Shuche Software, the developer behind open-source GPU cluster manager GPUStack, has raised ¥50 million in Seed funding. The round was led by Shuihua Weilai (水华未来), with participation from GRC Fuhua Capital (GRC富华资本).

What the Company Does

GPUStack is an open-source GPU cluster manager and enterprise AI infrastructure platform. It was built by a founding team drawn from the core team behind the Kubernetes management platform Rancher.

The platform manages heterogeneous GPU and NPU fleets spanning NVIDIA, AMD, Huawei Ascend, Hygon, Cambricon, Moore Threads and MetaX accelerators. It deploys and serves large models through inference engines such as vLLM, SGLang and TensorRT-LLM, exposing OpenAI-compatible APIs.

Traction

GPUStack was open-sourced in July 2024. According to the company, it has surpassed:

More than 110,000 cumulative users across 121 countries
Over 250 enterprise customers

Deal Summary

Company: GPUStack (Shuche Software)
Round: Seed
Amount: ¥50 million
Date: October 9, 2026
Lead Investor: Shuihua Weilai (水华未来)
Investors: GRC Fuhua Capital (GRC富华资本), Shuihua Weilai (水华未来)
Stage: Seed
Domain: gpustack.ai

Why It Matters

GPUStack addresses a core pain point in enterprise AI infrastructure: managing mixed accelerator fleets without locking customers into a single vendor. Its support for NVIDIA, AMD, Huawei Ascend, Hygon, Cambricon, Moore Threads and MetaX hardware positions it to serve organizations that need to run large-model inference across diverse and often constrained compute environments.

The founding team's background in Kubernetes management at Rancher is directly relevant to the orchestration problem GPUStack solves. Combined with an open-source distribution model that has already reached users in 121 countries and more than 250 enterprise customers, the company has built both developer mindshare and commercial traction at the Seed stage.

For acquirers and strategic investors in AI infrastructure, model-serving tooling, and GPU orchestration, GPUStack represents an early-stage asset with a broad hardware compatibility footprint and an OpenAI-compatible serving layer.