LANCUN Completes RMB 10 Million-Level Angel Round to Advance Multimodal Edge Embodied-AI Modules
LANCUN has completed an RMB 10 million-level angel round led by Xiaofanzhuo Capital, with participation from Shenran Fund and Tuojie Capital serving as the exclusive financial adviser for this and future rounds. The funding will support R&D and infrastructure for LANCUN’s next-generation embodied-AI foundation, helping connected devices move from interoperability to intelligent collaboration.
Source: 拓界交易
Positioning for a Trillion-Yuan Market
Securing a Core Gateway for AI-Native Hardware
Industry forecasts indicate that China’s AI hardware market, excluding smartphones and automobiles, will exceed RMB 1.1 trillion for the first time in 2025 and maintain rapid growth over the following five years. Demand for intelligent consumer products and public services is accelerating a fundamental shift in interaction—from person-to-person and human-machine interaction toward intelligent coordination among multiple devices.
LANCUN is focused on a core link in the AI hardware value chain: multimodal interactive intelligence modules. Rather than a single terminal product, the module is a standardized, platform-based embodied-AI foundation for toys, collectibles, smart homes, wearables, and industrial robots. By packaging affective computing, safety, multimodal perception, and autonomous decision-making into one system, LANCUN can rapidly add personality and intelligence to different devices while building a business model with strong network effects and platform expansion potential.
Affective Intelligence and Native Safety
Defining the Next Standard for Human-Machine Interaction
Many AI interaction solutions remain limited by weak emotional awareness, short-lived memory, and safety concerns. LANCUN’s core technology is designed to address both affective interaction and native safety at the same time.
Affective agent (EmoMonte): LANCUN developed China’s first end-to-end affective voice agent, capable of recognizing 50 human emotions and expressing 17 emotional states. Its simulation-and-pruning approach reduces interaction latency to less than 0.3 seconds and improves memory capacity by two orders of magnitude compared with competing solutions. By tightly coupling emotion with dialogue strategy, it brings greater warmth and continuity to human-machine interaction.
Native safety system: For demanding global compliance requirements, especially in children’s use cases, LANCUN developed China’s first AI-model-native safety and instruction-following system. Its four-layer architecture includes value anchoring and logical-consistency constraints, helping make AI behavior reliable, controllable, and trustworthy for commercialization at scale.
Cloud Optimization, Algorithm Innovation, and Hardware Cost Reduction
Continuously Expanding the Agent Ecosystem
LANCUN is advancing several R&D tracks in parallel around its edge-cloud-intelligence technology system.
Cloud: the company is optimizing the multi-agent collaboration network on its main platform to improve concurrent scheduling and intelligent task decomposition. Drawing on generative-agent architecture, the system enables agents to consolidate memory and refine their personas while not actively interacting, supporting continuous evolution.
Algorithms: information-entropy-density compression substantially reduces token transmission overhead during model inference, extending the system’s existing 70% cost optimization. Based on dual-process cognition (System 1/2), LANCUN is also building a coordinated decision mechanism that combines intuitive affective responses with rational reasoning, making agent decisions closer to human cognition.