DeepX has launched an EECP promotion where users can experience hardware and software firsthand. It is expected that AI chips that can be used in various fields of artificial intelligence such as autonomous vehicles, factory automation, AI-based security systems, and the AI server market will be pre-verified by approximately 40 global customers at home and abroad.

DeepX, an AI semiconductor source technology company, announced on the 4th that it has provided its flagship product 'DX-M1' to approximately 40 global customers at home and abroad in a pre-verification form before mass production.

DeepX is currently running an EECP (Early Engagement Customer Program) program, and is currently running a promotion that allows users to quickly experience the small camera module, a single-chip solution equipped with DX-V1, the M.2 module equipped with DX-M1, an AI accelerator solution, and DXNN, DeepX's developer environment.

Through this, customers receive DeepX’s hardware and software in a pre-verified form before mass production. By installing it in mass-produced products and receiving technical support from DeepX, AI technology can be applied to various embedded systems and objects.

Currently, DeepX's 'DX-M1' is being installed in mass production development products of approximately 40 global companies related to robots, smart mobility, AI video security systems, and AI servers, and is undergoing pre-verification testing.

The robotics and smart mobility sectors require technologies such as autonomous driving and perception, and the demand for low-power, high-performance AI semiconductors that can be mounted in small form factors is increasing. The physical security market is the market with the fastest AI penetration rate, and the realization of functions for AI semiconductor-based object detection and intelligent image analysis is essential.

DeepX boasted that its 'DX-M1', which uses a 5nm process, especially in the physical security market, has demonstrated overwhelming performance efficiency compared to power consumption, and has secured a significant gap compared to global competing technologies.

The DX-M1 supports real-time AI computation processing at over 30 FPS for multi-channel images of 16 or more channels with a single chip, and also has a wide spectrum of AI model support, from the YOLOv5 model, which is most commonly used in object recognition, to the latest YOLOv8 and the Vision Transformer model.

While competing AI semiconductors currently in the global market use 32MB to 50MB of cache memory, DeepX's DX-M1 product boasts that it is superior in key functions of AI semiconductors such as computational processing performance, AI computational accuracy, and types of supported AI algorithms while using only about a quarter of the cache memory.

The DX-M1 is provided in the form of a small M.2 module that can be easily and quickly connected to existing embedded systems, allowing customers to easily integrate low-power, high-performance AI solutions without modifying existing systems, making it easy to apply.

In recognition of this, the DX-M1 won innovation awards in both embedded technology and robotics categories at 'CES 2024', the world's largest IT and home appliance exhibition to be held in January, raising DeepX's profile worldwide. This year's Innovation Awards include products from global companies such as Samsung, Qualcomm, and MediaTek.

DeepX will unveil the DX-M1 at the CES 2024 exhibition in Las Vegas, USA in January 2024, through a standalone booth in the Las Vegas Convention Center North Hall. It said that it plans to introduce DeepX's solutions to potential customers and expand its business by attracting new customers and discovering new businesses.