NVIDIA, One Of AI’s Biggest Chipmakers, Uses AI To Design Faster Chips

(CTN News) – As the AI sector booms, companies are competing for Nvidia’s limited supply of GPUs. GPUs are used to train and develop AI applications. In order to meet the demands of the market, the chip giant is now using its own artificial intelligence to speed up its chips.

ChipNeMo, a system developed by Nvidia, aims to accelerate the production of GPUs.

GPU design is a labor-intensive process. According to Bryan Catanzaro, Nvidia’s vice president of applied deep learning research, it typically takes close to 1,000 people to build a chip. Each person needs to understand how the different parts of the design process interact.

ChipNeMo can assist in this regard. Using Meta’s Llama 2 as the foundation for the AI system, the company claims it trained the model with its own data. Catanzaro told the Wall Street Journal that ChipNeMo’s chatbot feature can answer questions about chip design, such as GPU architecture and generating chip design code.

As of now, the gains appear to be promising. In the months following ChipNeMo’s debut, Nvidia has found that the AI system has proven effective both in training junior engineers to design chips and in summarizing notes across 100 different teams.

Business Insider did not receive a response from Nvidia regarding whether ChipNeMo has resulted in faster chip production.

Nvidia is ramping up GPU production in order to gain a competitive advantage in the AI race. With its AI products such as its large language model Llama 2 and AI-powered Ray-Ban Smart Glasses, Meta plans to amass 600,000 GPUs by the end of 2024, including Nvidia’s A100s and other AI chips.

Nvidia’s commitment to building the best AI products appears to be paying off. The stock of the chip giant rose 4% on Monday to a record high, and Goldman Sachs expects the gains to continue through the first half of 2025.

The company is not the only one using artificial intelligence to speed up the semiconductor design process. As reported by the Wall Street Journal, Google’s DeepMind made an AI system that the company said would speed up the design process of its latest custom chip iteration.

The following month, Synopsys, a software giant, launched an artificial intelligence tool designed to boost the productivity of chip designers.

In addition, universities such as New York University are conducting research on the use of generative artificial intelligence (AI) to design chips more quickly.

SEE ALSO:

Snapchat Will Lay Off 10% Of Its Global Workforce