DEEPX launches DX-H1 V-NPU: 30W single-card solution that challenges GPU dominance

Ahead of CES 2026, DEEPX, a pioneer in ultra-low-power on-device AI semiconductors, announced the official launch of the DX-H1 vision neural processing unit (V-NPU), a dedicated video intelligence chipset capable of processing hundreds of AI video channels at just 30W.

Redefining video AI infrastructure
The DX-H1 V-NPU is an all-in-one video intelligence solution combining a multi-channel engine capable of decoding, encoding and transcoding with dedicated NPU architecture. Unlike traditional setups that require parallel configurations of GPU servers and hardware codecs, the DX-H1 V-NPU handles the entire pipeline on a single chip — from stream input to preprocessing, AI inference, and re-encoding.

The chip's launch represents a paradigm shift in data center architecture. By integrating video input, compression, and AI inference — processes typically requiring multiple GPU servers and separate codec hardware — onto a single card, the DX-H1 V-NPU is transforming video AI infrastructure's fundamental unit from the GPU into the V-NPU.

Unmatched efficiency and sustainability
According to DEEPX testing data, integrating these processes achieves approximately 80% savings in hardware costs and 85% savings in power consumption compared to GPU-based solutions for the same channel density, all while maintaining 24/7 real-time inference performance.

This efficiency offers a critical structural alternative for smart cities and industrial surveillance, addressing growing challenges such as data center power constraints, stringent ESG requirements, and GPU supply chain uncertainties.

"Large-scale video AI can no longer be a secondary task borrowing spare resources from general-purpose GPUs; it must evolve into a dedicated industry running on specialized chipsets," said Lokwon Kim, CEO of DEEPX. "The DX-H1 V-NPU is not merely a low-cost alternative but a fundamental redesign of video intelligence infrastructure, optimizing memory hierarchy and computation scheduling for environments where video streams pour in by the second."

From add-on to foundation
As the DX-H1 series has added AI capabilities to existing legacy systems such as CCTV and NVR, the new V-NPU variant builds on the series' success and is designed as a foundational platform. It targets new deployments in smart cities, traffic control centers, and large industrial complexes, centering the server architecture around the V-NPU rather than the GPU.