OpenAI is developing its first in-house AI chip in partnership with Broadcom to power its growing suite of AI tools, including ChatGPT, and future hardware products that will demand greater computing resources.
Under the partnership, Broadcom will handle the hardware design, while OpenAI focuses on optimizing the chip for its AI ecosystem. The new processors are expected to roll out in the second half of 2026, specifically tailored to meet OpenAI’s performance requirements rather than designed for long-term use.
Custom AI Chips: The Logical Next Step
OpenAI continues to rely heavily on Nvidia GPUs, and even holds the option to acquire a stake in the company. However, with global demand for custom AI chips surging, the collaboration with Broadcom marks a strategic move to reduce dependency and secure dedicated computing power. Broadcom’s stock saw an immediate boost following the announcement.
Beyond software, OpenAI is also venturing into hardware—having invested billions in a partnership with Jony Ive, the former Apple design chief, to create an AI-powered consumer device. While development continues, reports suggest that CEO Sam Altman and Ive are still debating the product’s identity and design direction.
Meanwhile, OpenAI is expanding its software capabilities with AgentKit, unveiled by Sam Altman at the company’s DevDay event. The new toolkit allows developers to build customized AI agents for specific tasks and workflows, signaling OpenAI’s ambition to become a central platform for AI-driven automation.
Although ChatGPT remains OpenAI’s flagship product, AI agents are rapidly emerging as the next frontier, reshaping how both individuals and businesses integrate artificial intelligence into everyday operations.
