OpenAI Builds Custom AI Chip with Broadcom: Should Nvidia Worry?

OpenAI, the creator of ChatGPT, is teaming up with Broadcom (AVGO) to develop a custom artificial intelligence chip, according to media reports. The move is designed to reduce OpenAI's reliance on Nvidia's (NVDA) high-priced GPUs, potentially challenging Nvidia's pricing power in the AI chip market.

Partnership Details

OpenAI and Broadcom are working together to design a custom AI chip for training and running large language models. Specifics on the timeline and investment remain undisclosed, but the project aims to improve cost efficiency and supply chain diversification.

Threat to Nvidia?

While the custom chip could reduce demand for Nvidia's GPUs in the long run, analysts argue that Nvidia's dominance is still formidable due to its superior performance and robust CUDA ecosystem. Moreover, OpenAI remains a relatively small customer compared to giants like Microsoft and Google.

Broader Context

This development comes as major tech companies race to develop their own chips to cut costs and boost efficiency. Google, Amazon, and Microsoft already have custom AI chips. However, building a competitive chip requires massive investment and deep engineering expertise.

What It Means for Investors

Nvidia investors should monitor the situation but not overreact. Nvidia's lead in AI chips is backed by strong performance and software moats. For Broadcom, the partnership strengthens its position in custom semiconductor design.