OpenAI and Broadcom to release chip in 2026

By Alimat Aliyeva

OpenAI is preparing to begin mass production of its first proprietary artificial intelligence (AI) chip in 2026, in partnership with Broadcom Inc., the Financial Times reported, citing multiple sources familiar with the collaboration, Azernews reports, citing foreign media.

Unlike many chip manufacturers, OpenAI plans to use this custom-designed chip exclusively for its own internal operations rather than selling it to external customers, according to an insider close to the project. The chip is being co-developed with Broadcom, a major player in semiconductor technology, aiming to optimize performance specifically for AI workloads.

This strategic move reflects OpenAI’s efforts to address the skyrocketing demand for computing power driven by the rapid growth of AI applications like ChatGPT. It also signals the company’s intention to reduce its heavy reliance on Nvidia’s hardware, which currently dominates the AI processing market.

Industry experts see this development as a significant step toward vertical integration in the AI sector, allowing OpenAI greater control over its technology stack and potentially lowering costs. It also underscores the growing trend of tech companies designing custom chips tailored to their unique needs, similar to moves by giants like Google and Apple.

With AI models becoming increasingly complex and resource-intensive, having a dedicated, highly efficient chip could give OpenAI a competitive edge in both performance and scalability.