ArcheFlowX

, trading as ArcheFlowX, provides AI inference chips and real-time inference infrastructure for the artificial intelligence and semiconductor industries. The company focuses on developing proprietary LPU+ architecture for low-latency, high-efficiency AI inference across cloud, edge, and endpoint computing scenarios. The company was founded by Bin Yang, who has more than 20 years of chip and system architecture experience and previously helped build Huawei’s processor team in the US before leading wireless baseband algorithm and chip-related work. The company works with cloud service providers, computing centers, AI infrastructure operators, industrial customers, and edge device manufacturers to provide AI inference computing products and infrastructure solutions. Its core technology is the LPU+ architecture, an AI inference architecture built around spatial-temporal compiler technology, hardware dataflow architecture, and high-bandwidth memory. The platform is designed to improve inference latency, execution determinism, energy efficiency, and deployment efficiency for large language models, multimodal AI, mixture-of-experts models, and Agentic AI workloads. The company’s revenue is generated through the commercialization of AI inference chips, LPU+ computing products, and inference infrastructure solutions supplied to cloud vendors, computing center operators, AI companies, and edge intelligence customers.

1. Angel

   Jan 20, 2026

   Amount

   Undisclosed

   Investors

   Locked