Silicon Flow Raises Nearly 2.9 Billion RMB in 2026 Financing Rounds
Chinese AI infrastructure startup Silicon Flow completed its Series B+ (phase 2) and Series C funding rounds, bringing its total equity financing in 2026 to nearly 2.9 billion RMB. The company will use the capital to boost R&D in inference engine optimization, heterogeneous compute scheduling, model-chip adaptation, and global market expansion. As enterprise demand for deploying generative AI models rapidly expands, reducing computing costs through software-hardware optimization has become a critical operational priority. Silicon Flow's substantial funding highlights the growing strategic importance of specialized inference platforms and multi-chip scheduling in the AI ecosystem. The financing round was backed by a diverse syndicate of state-level funds, central and local state-owned capital, tech industry investors, and venture firms. Founded in August 2023, Silicon Flow offers cloud API services, enterprise Model-as-a-Service (MaaS) platforms, dedicated instances, and integrated AI appliances tailored for LLMs, image generation, and video generation workloads.
## BACKGROUND
LLM inference engine optimization uses techniques such as continuous batching and quantization to reduce memory usage and accelerate text generation speed. Heterogeneous compute scheduling manages dynamic workload distribution across diverse processor types, such as different GPU architectures and AI accelerators. Model-as-a-Service (MaaS) is a delivery model that provides cloud-based API access to pre-trained machine learning models, allowing organizations to deploy AI applications without managing complex infrastructure.