Winbond Unveils Powerful Edge AI Computing Technology

Winbond Electronics Corporation, a leading global supplier of semiconductor memory solutions, has introduced a groundbreaking technology for affordable Edge AI computing in mainstream applications. The company's new customized ultra-bandwidth elements (CUBE) optimize memory technology for seamless performance in generative AI on hybrid edge/cloud applications.

CUBE enhances the performance of front-end 3D structures such as chip on wafer (CoW) and wafer on wafer (WoW), as well as back-end 2.5D/3D chip on Si-interposer on substrate and fan-out solutions. It is designed to meet the growing demands of edge AI computing devices and is compatible with memory density from 256 Mb to 8 Gb with a single die. Additionally, it can be 3D stacked to enhance bandwidth while reducing data transfer power consumption. Winbond's CUBE technology enables seamless deployment across various platforms and interfaces and is well-suited for advanced applications such as wearable and edge server devices, surveillance equipment, ADAS, and co-robots.

"The CUBE architecture enables a paradigm shift in AI deployment," says Winbond. "We believe that the integration of cloud AI and powerful edge AI will define the next phase of AI development. With CUBE, we are unlocking new possibilities and paving the way for improved memory performance and cost optimization on powerful Edge AI devices."

Key Features of CUBE:

  • Power efficiency: CUBE consumes less than 1pJ/bit, ensuring extended operation and optimized energy usage.
  • Superior performance: CUBE offers bandwidth capabilities ranging from 32 GB/s to 256 GB/s per die, exceeding industry standards.
  • Compact size: CUBE provides a range of memory capacities from 256Mb to 8Gb per die, allowing it to fit seamlessly into smaller form factors. Through-silicon vias (TSVs) enhance performance, improve signal and power integrity, reduce IO area, and aid in heat dissipation.
  • Cost-effective solution with high bandwidth: CUBE's IO achieves a data rate of up to 2 Gbps with a total of 1K IO, delivering outstanding cost-effectiveness. When combined with legacy foundry processes like 28 nm/22 nm SoC, CUBE achieves ultra-high bandwidth capabilities.
  • Reduction in SoC die size for improved cost efficiency: By stacking the SoC (top die without TSV) on top of the CUBE (bottom die with TSV), the SoC die size can be minimized, resulting in cost advantages and improved efficiency for Edge AI devices.

"CUBE can unleash the full potential of hybrid edge/cloud AI to elevate system capabilities, response time, and energy efficiency," added Winbond. "Winbond's commitment to innovation and collaboration will enable developers and enterprises to drive advancement across various industries."

Winbond is actively collaborating with partner companies to establish the 3DCaaS platform, leveraging CUBE's capabilities. By integrating CUBE with existing technologies, Winbond aims to provide cutting-edge solutions that empower businesses to thrive in the era of AI-driven transformation.