Suanova Technology Limited (“Suanova”) has entered into a comprehensive strategic cooperation agreement with InfiX.ai. The collaboration will focus initially on healthcare, using Suanova’s domestically developed Shanghai Cube high-density computing infrastructure to advance deployment of training and inference integrated AI platforms with continuous self-learning capabilities.
Under the agreement, Suanova will supply core computing power and foundational support, while InfiX.ai will contribute its training, inference and multimodal AI capabilities to create a unified hardware-software stack designed for domain-specific medical applications.
Both parties have already worked with leading medical institutions to conduct clinical validation in key areas including cancer GenAI, foundational medical large language models (LLMs), and personalised cancer treatment planning. The companies said the new programme builds on this groundwork and is intended to move from validation to wider real-world deployment.
A central element of the initiative is the Shanghai Cube architecture, which uses a high-density design featuring 128 GPUs per rack supported by liquid cooling.
The platform is described as the highest-density domestically developed GPU supernode product, and it integrates a set of domestic core components, including liquid-cooling systems, high-performance parallel storage, retimers and motherboard capacitors.
By packaging these elements in a one-stop solution for large-scale model and system deployment, Shanghai Cube is positioned to improve efficiency for enterprises scaling compute-intensive AI workloads.
InfiX.ai, meanwhile, is said to be developing a Decentralized Co-GenAI Network connecting computing power, models, platforms and intelligent applications, with the aim of enabling organisations to train, deploy and own domain-specific AI based on proprietary data, expertise and workflows.
InfiX.ai’s platform is designed to support local execution of the AI workflow—from training and fine-tuning through to inference—an approach intended to strengthen data security and help meet stringent regulatory expectations in healthcare and other sensitive sectors.
The companies also highlighted technical benefits from the integration, including reduced memory usage and lower computing resource requirements, enabling higher throughput and support for larger-scale models on equivalent hardware.
The platform is further described as capable of capturing data for incremental training and integrating user feedback for fine-tuning and reinforcement learning, allowing performance to evolve in line with changing operational needs.
Commenting on the deal, Daliang Chen, CEO of Suanova, said: “This partnership with InfiX.ai represents an important milestone in Suanova’s expansion into medical AI. Leveraging the Shanghai Cube high-density domestic computing platform, we aim to accelerate the adoption of medical AI in real-world clinical settings.
"This collaboration not only brings together the complementary strengths of both companies from a technological perspective, but also serves as a key step in advancing the domestic computing ecosystem. Looking ahead, we will continue to work closely with our partners to drive the deep integration of artificial intelligence across diverse industries.” Daliang Chen
