Broadcom Secures $10 Billion Deal with OpenAI for Custom AI Chips

OpenAI is set to mass-produce its first proprietary AI chip with Broadcom, marking a shift away from Nvidia. The chip will be used internally for ChatGPT's 700 million weekly users and is expected to boost Broadcom's chip shipments strongly starting next year. The partnership began last year, and OpenAI plans to use the chips rather than making them available to external customers. The custom chip development signals a potential shift in the AI hardware landscape, where Nvidia has maintained dominance.

OpenAI, the company behind the popular AI chatbot ChatGPT, is set to produce its first proprietary AI chip in partnership with Broadcom, marking a significant shift away from its reliance on Nvidia. The chip, expected to be mass-produced starting next year, will be used internally to support ChatGPT's 700 million weekly users. This move signals a potential shift in the AI hardware landscape, where Nvidia has maintained dominance.

The partnership between OpenAI and Broadcom began last year, with OpenAI planning to use the chips internally rather than making them available to external customers. The custom chip development is expected to boost Broadcom's chip shipments significantly starting next year. OpenAI's move to reduce its reliance on Nvidia comes amidst growing demand for AI hardware and increasing costs associated with Nvidia's accelerators.

Broadcom's CEO, Hock Tan, confirmed during an investor call that the company expects artificial intelligence revenue growth for fiscal 2026 to "improve significantly" after securing more than $10 billion in AI infrastructure orders from new customers. The partnership with OpenAI is a key driver of this growth, with the first units expected to be shipped as early as next year.

The development of OpenAI's in-house AI chip also aligns with its broader strategy to diversify its chip supply and reduce costs. The company has been exploring various options to achieve this, including working with Taiwan Semiconductor Manufacturing Co. (TSMC) and AMD to develop custom chips.

OpenAI's move to produce its own AI chip is part of a broader trend among tech giants to build custom chips to handle AI workloads. Companies like Google, Amazon, and Meta have also developed custom chips to meet the surge in demand for computing power to train and operate AI models.

The development of OpenAI's in-house AI chip is a significant milestone in the company's history and could have far-reaching implications for the AI hardware landscape. However, it remains to be seen how this move will impact Nvidia's dominance in the market and the broader AI ecosystem.

References:
[1] https://www.livemint.com/technology/tech-news/openai-to-build-own-ai-chips-with-broadcom-marking-bold-shift-from-nvidia-reliance-report-11757035663266.html
[2] https://www.reuters.com/business/openai-launch-its-first-ai-chip-2026-with-broadcom-ft-reports-2025-09-05/