Nvidia Could Possibly Tap Into TSMC's CoWoS Packaging For Next-Generation GPUs

[Russ ;x]

According to the new DigiTimes report, Nvidia will be among the three major customers that will be packing CoMoS from TSMC this year, with the other two being Xilinx and HiSilicon.

CoWoS (Chip-on-Wafer-on-substrate) is a 2.5D encapsulation technology that integrates multiple segments into a single interferer. The list of basic benefits includes much less space, increased bandwidth and lower power consumption.

Speaking of CoWoS, TSMC jointly developed with Broadcom recently a massive 1700 mm2 interferer. In short, the concept consists of several interconnected interconnecting interconnects on a single chip. The new design is emerging in the Broadcom HPC (High Performance Computing) processor based on TSMC's EUV 5nm (N5) node. An enhanced CoWoS platform can die many System Templates (SoC) on the system and up to six stacks of High Bandwidth Memory (HBM), and the total mass memory reaches 96 GB. TSMC is promoting a memory bandwidth of up to 2.7 terabytes per second, 2.7 times faster than the company's previous CoWoS system that started in 2016.

Coming back to Nvidia, the chipmaker is no stranger to CoWoS packaging. The technology is present on a number of high-end Titan, Quadro and Tesla graphics cards, dating back to the Pascal era. The GP100 (Pascal) and GV100 (Volta) silicons come out of TSMC's furnace and use CoWoS packaging. The first is based on the foundry's 16nm FinFET manufacturing process, while the latter profits from the 12nm node.

AMD's Vega 20 7nm silicon is also packaged with CoWoS. DigiTimes did not include AMD on its top three list, which means Nvidia, Xilinx and HiSilicon will ultimately gobble up TSMC's CoWoS production capacity. The foundry will reportedly pump out between 6,000 to 8,000 wafers per month so there should be enough product to go around.

In any event, we don't expect Nvidia's consumer graphics cards to feature CoWoS technology due to the high costs that it would post to the consumer. The chipmaker will probably turn to CoWoS for its Quadro and Tesla offerings like in the past. Ampere is the codename for Nvidia's next-generation GPU architecture, but there hasn't been an indication that it will utilize a MCM (multi-chip module) design, though. Hopper, on the other hand, might rely on MCM.

We've already seen a trio of mysterious Nvidia graphics cards sporting ginormous amounts of CUDA cores and memory. Although we're unsure of their nature, the unidentified graphics cards should make use of one of Nvidia's upcoming GPU architectures.