DEEPX Wins 2026 Patent Technology Award for Pipeline-Based AI Accelerator

What happened

On Nov. 11, DEEPX announced it received the Hong Dae-yong Award at the 2026 Patent Technology Awards for its patent titled 'Computing Acceleration Device for Artificial Neural Networks Having a Pipeline Structure' (Patent No. 10-2263598). The patented architecture divides the AI accelerator's computing process into six combinational logic circuit stages connected by temporary storage buffers, improving processing speed and power efficiency by reducing the storage and movement of intermediate data generated during AI computations. In a comparison with a conventional design using the same Xilinx XC2VP70 FPGA, computational performance increased 12-fold and memory consumption fell to approximately one-341st of the previous level. Founded in 2018, DEEPX has filed 606 patent applications to date, of which 186 have been registered; its 600th application relates to DX-Newton, an API layer that lets developers retain an Nvidia Isaac-based robotics development environment while replacing only the AI inference acceleration component with a DEEPX NPU. The source does not state the cause.

Impact and signal

Affected components and suppliers

Source and time

Source: The Elec · 11/12/2026 · View original source