CoWoS®

The CoWoS®-S (Chip on Wafer on Substrate with silicon interposer) platform provides best-in-class package technology for ultra-high performance computing applications, such as artificial intelligence (AI) and super-computing. This wafer level system integration platform offers high density interconnects and deep trench capacitors over a large silicon interposer area to accommodate various functional top die/dice, including logic chiplets, with High Bandwidth Memory (HBM) cubes stacked over it. Currently, an interposer up to 3.3X-reticle size (or ~2700mm2) is ready for production. The CoWoS®-L and CoWoS®-R platform are recommended for larger than 3.3X-reticle interposer sizes. Different interconnect options provide more flexibility integration to meet performance target.

CoWoS<sup>®</sup>-S

CoWoS®-R (Chip on Wafer on Substrate with silicon interposer with fan-out RDL interposer) is a member of CoWoS® advanced packaging family that leverages a redistribution layer (RDL) interposer as the interconnect between System on Chip (SoC) and/or high bandwidth memory (HBM) to achieve heterogeneous integration. An RDL interposer is comprised of polymer and copper traces and is relatively flexible. This enhances C4 joint integrity and allows the package to scale its size to meet very complex functional demands.

CoWoS<sup>®</sup>-R

Key CoWoS®-R features include:

  1. An RDL interposer of up to 6 copper layers for routing with a minimum 4μm pitch (2μm line width/spacing).

  2. The RDL interconnect offers good signal and power integrity with a lower RC value routing line to achieve a high data transmission rate. The co-planar Ground-Signal-Ground-Signal-Ground (GSGSG) and interlayer ground shielding with six RDL interconnections offers superior electrical performance.

  3. RDL layer and C4/underfill (UF) layers provide an excellent buffer thanks to the coefficient of thermal expansion (CTE) mismatch between an SoC and the corresponding substrate. The strain energy density is greatly reduced in C4 bump area.

    CoWoS<sup>®</sup>-R

CoWoS®-L is one of the chip-last packages on the CoWoS® (Chip on Wafer on Substrate) platform. It combines the merits of CoWoS®-S and InFO (Integrated Fan-Out) technologies to provide the most flexible integration using an interposer with a Local Silicon Interconnect (LSI) chip for die-to-die interconnect and RDL layers for power and signal delivery.

Key CoWoS®-L features include:

  1. LSI chips for a high routing density die-to-die interconnect through multiple layers of sub-micron copper lines. LSI chips can feature of a variety of connection architectures, e.g., System on Chip (SoC)-to-SoC, SoC-to-chiplet, SoC-to-High Bandwidth Memory, within each product, and can be used repeatedly in multiple products. The corresponding metal types, layer counts, and pitches align with the offering from CoWoS®-S.
  2. A molding-based interposer with a wide pitch of RDL layers on the front-side, back-side and Through InFO Via (TIV) that delivers signal and power provides a low loss of high frequency signals during high-speed transmission.
  3. The ability to integrate additional elements, such as stand-alone embedded Deep Trench Capacitors, underneath the SoC die to improve power management.
CoWoS<sup>®</sup>-L
The Chronicle of CoWoS
  • 2024
    • News Release
      2024/01/10
      GUC Taped Out UCIe 32G IP using TSMC's 3nm and CoWoS Technology
      LEARN MORE
  • 2023
    • News Release
      2023/04/06
      GUC Taped Out 3nm 8.6Gbps HBM3 and 5Tbps/mm GLink-2.5D IP using TSMC Advanced Packaging Technology
      LEARN MORE
  • 2022
    • News Release
      2022/07/07
      GUC Demonstrate World's First HBM3 PHY, Controller, and CoWoS Platform at 7.2 Gbps
      LEARN MORE
  • 2021
    • Industry Publication
      ECTC 2021
      Wafer Level System Integration of the Fifth Generation CoWoS®-S with High Performance Si Interposer at 2500 mm2
      LEARN MORE
    • News Release
      2021/08/31
      GUC Announces Industry Highest Bandwidth and Power Efficient Die-to-Die (GLink 2.0) Total Solution
      LEARN MORE
      2021/06/08
      GUC Tapes Out AI/HPC/Networking Platform on TSMC CoWoS® Technology Validating 7.2 Gbps HBM3 Controller and PHY, GLink-2.5D and 112G-LR SerDes IPs
      LEARN MORE
  • 2020
    • Industry Publication
      ECTC 2020
      Design and Analysis of Logic-HBM2E Power Delivery System on CoWoS® Platform with Deep Trench Capacitor
      LEARN MORE
    • News Release
      2020/11/17
      GUC Die-to-Die (D2D) Total Solution Opening the New Era of Flagship SoC
      LEARN MORE
      2020/09/09
      Rambus Advances HBM2E Performance to 4.0 Gbps for AI/ML Training Applications
      LEARN MORE
      2020/03/03
      TSMC and Broadcom Enhance the CoWoS Platform with World’s First 2X Reticle Size Interposer
      LEARN MORE
  • 2019
    • Industry Publication
      IEDM 2019
      Integrated Deep Trench Capacitor in Si Interposer for CoWoS Heterogeneous Integration
      LEARN MORE
    • Production Milestone
      More than 60 product tape-outs are in production or in development as of Aug. 2019
    • Customer Product
      Industry 1st 7nm GPU w/ deep learning accelerator
      • 1TB/s in BW
      • 4 HBM2
      • 1X reticle interposer
      AI training accelerator w/ 1.2TB/s in BW
      • N16+
      • 4 HBM2
      • 1.5X reticle interposer
  • 2018
    • Customer Product
      Fujitsu A64FX
      LEARN MORE
      Build your 56G Enterprise Networking ASICs with MediaTek
      LEARN MORE
  • 2017
    • Customer Product
      Broadcom Announces Industry's First Silicon-Proven 7nm IP for ASICs in Deep Learning and Networking Applications
      LEARN MORE
      NEC "Aurora" Vector Engine vector processor
      LEARN MORE
      Nvidia TESLA GV100
      LEARN MORE
    • Industry Publication
      IEEE Transaction on electronics devices 2017
      Wafer-Level Integration of an Advanced Logic-Memory System Through the Second-Generation CoWoS Technology
      LEARN MORE
      IEDM 2017
      Advanced Heterogeneous Integration Technology Trend for Cloud and Edge
      LEARN MORE
  • 2016
    • Customer Product
      Nvidia TESLA GP100
      LEARN MORE
      Industry's first ASIC based AI accelerator from learning only to learning+inference
    • Industry Publication
      SEMICON Taiwan 2016
      Interposer Technology: Past, Now, and Future
      LEARN MORE
      SEMICON Taiwan 2016
      WLSI Extends Si Processing and Supports Moore's Law
      LEARN MORE
  • 2015
  • 2014
    • Customer Product
      HiSilicon Hi1616
      LEARN MORE
    • Industry Publication
      IEDM 2014
      A manufacturable interposer MIM decoupling capacitor with robust thin high-K dielectric for heterogeneous 3D IC CoWoS wafer level system integration
      LEARN MORE
      IEDM 2014
      Wafer Level System Integration for SiP
      LEARN MORE
      CICC 2014
      New System-in-Package (SiP) Integration technologies
      LEARN MORE
  • 2013
    • Industry Publication
      VLSI 2013
      Manufacturability Optimization and Design Validation Studies for FPGA-Based, 3D Integrated Circuits
      LEARN MORE
      IITC 2013
      Innovative Wafer-based Interconnect Enabling System Integration and Semiconductor Paradigm Shifts
      LEARN MORE
      ECTC 2013
      Reliability Evaluation of a CoWoS-enabled 3D IC Package
      LEARN MORE
  • 2012
    • Customer Product
      Xilinx 7V2000T/7V580T
      LEARN MORE
    • Industry Publication
      VLSI 2012
      An ultra-thin interposer utilizing 3D TSV technology
      LEARN MORE
  • 2011
    • Industry Publication
      ECTC 2011
      Advanced Reliability Study of TSV Interposers and Interconnects for the 28nm Technology FPGA
      LEARN MORE

Related Information