TSMC Kicks 2nm Into High Gear To Fulfil Demands From NVIDIA & Apple

Low Boon Shen
2 Min Read
TSMC Kicks 2nm Into High Gear To Fulfil Demands From NVIDIA & Apple

TSMC Kicks 2nm Into High Gear To Fulfil Demands From NVIDIA & Apple

TSMC Kicks 2nm Into High Gear To Fulfil Demands From NVIDIA & Apple

TSMC Kicks 2nm Into High Gear To Fulfil Demands From NVIDIA & Apple

To keep itself competitive against rival foundries such as Intel and Samsung, Taiwanese foundry TSMC has reportedly started pre-production work for 2nm trial production, in time for 2025 mass production in order to serve two of its major customers, Apple and NVIDIA, according to Economic Daily (UDN).

Sources from UDN claimed that engineers and support staff are now on the Zhuke R&D plant in Baoshan, Taiwan; and the facility is set to churn out 1,000 wafers this year, with trial production set for 2024 and mass production on the year after. The production is set to kick off in Fab 20, which is still under construction.

TSMC Kicks 2nm Into High Gear To Fulfil Demands From NVIDIA & Apple 6

TSMC Kicks 2nm Into High Gear To Fulfil Demands From NVIDIA & Apple 6

NVIDIA for their part has introduced AI into the mix – dubbed AutoDMP, this AI can calculate the best balance between cost, performance and power consumption – all while significantly reducing the compute needed to generate “floor-planning”, saving carbon footprint in the process.

TSMC’s 2nm node will be the first from the company to introduce GAAFET – which is a generation later than Samsung which will integrate GAAFET (MBCFET in their implementation) into 3nm-class nodes, while Intel is set to use another implementation called RibbonFET starting with Intel 20A node.

Source: Tom’s Hardware

Pokdepinion: The race is on – though it’d take lots of effort to draw TSMC’s customers away given how dominant their position has been since the FinFET era. 

Share This Article