Neuchips Joins Vecow and GSH to Enhance AI Data Processing

2025-03-06 smart

Hsinchu, Thursday, 6 March 2025.
The collaboration introduces a power-efficient AI solution to privately and securely enhance SQL data processing with offline models, significantly increasing large language model accessibility and efficiency.

Revolutionary Partnership Announcement

On March 5, 2025, Neuchips, a leading provider of AI Application-Specific Integrated Circuits (ASICs), announced a groundbreaking collaboration with Vecow and Golden Smart Home (GSH) Technology Corp. [1][2]. The partnership aims to revolutionize SQL data processing through a secure, private, and energy-efficient AI solution that enables real-time insights from enterprise databases using natural language queries [3].

Technical Innovation and Specifications

The solution combines Vecow’s ECX-3100 RAG Edge AI Inference Workstation with Neuchips’ innovative Viper series Gen AI card, achieving remarkable efficiency metrics. The Viper card supports a 12 billion parameter model while consuming only 45W of power [4]. The system features 32GB of additional memory capacity and native support for BF16 Structured Language Model, enabling it to offload over 90% of generative AI processing resources from the CPU [5].

Industry Impact and Implementation

Executive Vice President of Vecow, Joseph Huang, emphasizes the growing demand for on-premise generative AI applications and multimodal large language models (LLMs) [1]. The solution specifically addresses the challenge of complex database management by enabling natural language to SQL query conversion, significantly reducing the need for specialized SQL expertise [2]. The partnership will showcase this technology at Embedded World 2025 in Nuremberg, Germany, from March 11-13, at Vecow’s booth #3-449 [6].

Future Developments

Looking ahead to 2026, Neuchips has announced plans to focus on developing low-power multimodal ASICs to achieve even greater performance gains [3]. This strategic direction aligns with the industry’s growing demand for more efficient and powerful AI processing solutions, particularly in enterprise environments requiring secure, offline processing capabilities [4].

sources

offline AI data processing