OpenAI is developing its own AI chip, set for production next year, with a design finalization in the near future. This shift aims to decrease reliance on Nvidia chips for AI model training. Taiwan Semiconductor Manufacturing Co (TSMC) will fabricate the chip employing advanced 3-nanometer technology, integrating high-bandwidth memory and strong networking capabilities. Initially, the chip will be deployed on a limited scale to support AI model running, with future iterations expected to include enhancements. The initiative reflects ongoing trends in the industry, as tech companies heavily invest in AI infrastructure.
OpenAI plans to finalize its in-house AI chip design in the coming months, shifting away from reliance on Nvidia's chips for AI model training.
The chip will be manufactured by TSMC using advanced 3-nanometer technology, featuring high-bandwidth memory and extensive networking capabilities.
Collection
[
|
...
]