OpenAI’s AI Chip: What You Need to Know

On Track for Launch Next Year

According to a recent report from Reuters, OpenAI is moving forward with plans to develop its in-house AI chip, with a projected launch date of next year. Sources close to the project indicate that OpenAI is set to finalize its design over the next few months, with Taiwan Semiconductor Manufacturing Co (TSMC) handling the fabrication process.

Why an In-House Chip?

By creating its own AI chip, OpenAI aims to reduce its reliance on Nvidia’s chips for training and running AI models. The custom chip will utilize TSMC’s 3-nanometer technology, boasting high-bandwidth memory and extensive networking capabilities.

Initial Deployment and Future Plans

At launch, OpenAI plans to deploy its in-house chip on a limited scale, primarily using it for running AI models. The company also intends to develop future versions of the chip with more advanced processors and capabilities.

A Brief History of OpenAI’s Chip Development

This latest development follows a report from last year, which suggested OpenAI is collaborating with Broadcom to create a custom chip. OpenAI’s chip design team, led by former Google TPU engineer Richard Ho, has grown from 20 to 40 members in recent months as efforts ramp up.