Next-Generation Nvidia GPUs to Use TSMC’s CoWoS Packaging

Nvidia’s next-generation GPUs will most likely tap into TSMC’s CoWoS packaging in 2020. A report from DigiTimes claims that Nvidia will be one of three major clients to take advantage of this TSMC and CoWos technology.

Joining Nvidia are Xilinx and HiSilicon in being the others to implement CoWoS (Chip-on-Wafer-on-Substrate) packaging. CoWoS uses 2.5D packaging technology which features multiple chiplets on a single interposer. This leaves multiple advantages like power consumption, a smaller footprint, and increased bandwidth.

Nvidia’s main rival, Advanced Micro Devices (AMD) name is not present, despite the Vega 20 7nm silicon which packs CoWoS. With this, NVidia is at a competitive advantage as it will produce 6,000 to 8,000 wafers per months.

Further confirming, Nvidia, Xilinx and HiSilicon will be taking up most of the CoWoS production capacity. It’s unlikely that CoWoS technology is going to feature in Nvidia’s consumer-grade graphics cards due to its high cost. We may see it on other products emphasizing on data systems and high-performance computing (HPC)

Rumours state the next-generation Nvidia GPU architecture, codenamed “Ampere”, will utilize an MCM (multi-chip module) design. This leaves us in speculation that Nvidia will use CoWoS for its Quadro and Tesla offerings mirroring the past.

Nvidia has used CoWoS packaging before, most notably in their high-end Titan, Quadro and Tesla graphics cards. The GP100 (Pascal) is based on TSMC’s 16nm FinFET manufacturing process. While the GV100 (Volta) benefits from the 12nm node.

TSMC has recently teamed up with Broadcom to develop an interposer measuring in at a staggering 1,700 mm². They’ve achieved this feat by binding multiple interconnected interposers together on a solo wafer. The Broadcom HPC processor is going to be featuring this design when it comes out.

The processor will be based on TSMC’s EUV 5nm (N5) node. The new CoWoS platform totals up to 96GB of memory. It does this by housing multiple system-on-chip (SoC) dies, and six stacks of HBM (high-bandwidth memory).

Coming back to Nvidia, rumours suggest the company will be working on three new graphics cards with a whopping 7,936 CUDA Cores. These cards were apparently benchmarked with Geekbench 5 but the company is yet to make any specific announcements about the matter.

Thus these benchmarks may invoke some sense of scepticism. This still does not give us better detail on what NVidia has in store for us. Regardless, it’s not unforeseeable that future graphics cards from Nvidia are likely to use the new GPU architectures.