LeapFive News
A New Breakthrough in RISC-V Computing! Yueshiang AMC Brings DeepSeek’s Large Language Model Right into Your Pocket.
Release time:
2025-03-15
From “cloud-based queuing” to “having a personal co-pilot”—edge AI can do it all!
Is the DeepSeek Q&A response timing out again?
“No data center, no server—can a single USB drive really power the DeepSeek large language model?” Yes, we’ve made it happen!
Yue Fang Technology , as a leading domestic RISC-V architecture SoC chip company , successfully will Local Deployment of DeepSeek Large Models , to The USB drive form factor is redefining how AI computing is used.
AMC (AI Mini Computer) Powered by Yuefang Technology Self-developed RISC-V chip Built for direct USB connection to your host device, it’s compatible with multiple operating systems and supports the deployment of localized AI inference applications—such as vision-based applications powered by traditional neural networks. Now, AMC has further enhanced its capabilities with large-model applications, enabling cloud-independent operation that’s immune to network fluctuations. It’s plug-and-play, ready for instant inference at any time. Transform your ordinary PC into an AI workstation in seconds, keeping AI always on standby and ready to respond swiftly—becoming your personal AI co-pilot!
*AMC Live Shot | Your Personal Smart Assistant in Your Pocket
A small body holds great opportunities—moving from “usable” to “easy to use.”
🔵 Plug and play No server required—just plug in via USB and complete deployment in 10 seconds.
🔵 3W ultra-low power consumption ， Both computational performance and the model itself have room for optimization.
🔵 Local private knowledge base Protect data privacy, carry it with you, and keep it secure and under control.
🔵 Web page localization access (Beta version): Compared to queuing in the cloud and waiting for responses, your AI assistant is always online.
*Models that have already been successfully deployed and are in operation
UI interface: The design is simple and intuitive, allowing users to get started quickly without any complicated configurations—less is more.
OTA + Scenario-Based Models: Pioneering a New Paradigm for Edge AI Computing!
As AI computing moves toward multi-scenario applications today, Developers hope to train and deploy models with greater flexibility, while users look forward to AI that can precisely meet their personalized needs. However, traditional AI computing relies on the cloud, and issues such as high computational costs, network latency, and data privacy make it difficult to truly deploy AI in diverse edge applications.
The process of deploying DeepSeek to AMC has also prompted us to engage in deeper reflection and led to the following discoveries: DeepSeek’s miniaturization, low-cost, and edge-device OTA (Over-the-Air) upgrade capabilities have the potential to become a brand-new approach for deploying edge AI computing applications. Drive AI computing from the cloud to the edge, enabling AI to adapt more precisely and flexibly to localized applications.
Conceptual diagram
From “large and comprehensive” to “small and refined,” we’ve seen AMC’s potential in the AI computing field, which holds promise for making AI computing more precise and more efficient. Scenario-based models and OTA updates will enable developers to build AI applications faster and allow users greater freedom in their AI usage, driving AI computing toward a shift from general-purpose to customized solutions, from fixed to flexible architectures, and from single-use to multi-faceted applications—truly empowering countless industries and sectors.
Yuefang Technology, leveraging the RISC-V computing architecture, continuously explores innovative integrations of chips, edge AI computing, and intelligent connectivity, accelerating the advancement of domestically developed, independent AI computing. We believe that in the future, AI computing will no longer be an expensive resource but rather a universally accessible capability. Personalized intelligence is the core driving force propelling the world toward a smarter future!
AMC is now open for trials and welcomes all kinds of business discussions.
📞 Consult now to get a personalized AI computing solution!
Phone: +86-17727560859 (direct line to professional team)
*Some features mentioned in this article are still under development and optimization at various stages: features such as web page loading speed and lightweight AI model adaptation are currently being refined and will continue to be upgraded and improved. Meanwhile, knowledge base storage and the OTA subscription model remain at the concept-validation stage; in the future, we will gradually explore their practical implementation based on technological advancements. Currently, AMC is adapted to lightweight AI models, while full-scale large-model inference capabilities are still under exploration. We welcome developers and users alike to join us in testing and providing feedback, helping to drive the continuous evolution of the AI computing ecosystem.
