OpenAI plans to launch its AI chip in 2026 in collaboration with the global semiconductor leader, Broadcom. With this step, OpenAI aims to reduce the demand and dependency for computing power on Nvidia, while fulfilling the AI workloads on its own.
Although no official statement has been released by either OpenAI or Broadcom about the collaboration, according to the internal reports and the Financial Times’ coverage, the launch of the AI chip is scheduled for 2026. Alongside that, Broadcom’s CEO, Hock Tan, revealed a mystery customer investment of $10 billion in orders.
Tan stated, “One of these prospects released production orders to Broadcom, and we have accordingly characterised them as a qualified customer for XPUs.” Another source suggests that OpenAI plans to use the chip exclusively for internal purposes, without making it available to external customers.
OpenAI has taken this step to design its own chip, following Google, Meta, and Amazon’s initiative to deploy their in-house chip for model training and AI workload management.
OpenAI’s Collaboration with Broadcom-
OpenAI started talks with Broadcom last year; however, the plan for chip design and production has been obscure. Hock Tan revealed that Broadcom closed a customer deal for custom AI chip development the previous year, which happened to be the fourth major deal for the firm. Though Broadcom didn’t specify its client to be OpenAI, insiders close to the deal confirmed the rumours.
Tan further remarked that the deal helped Broadcom achieve growth, increasing demand, which will allow them to process the shipment of the chips by next year. Notably, the collaboration of OpenAI and Broadcom will allow the former to expand its computing power and the latter to sustain its dominance as one of the leaders in the AI hardware sector.
Why OpenAI is Developing its Own Chip?
OpenAI is making generative AI more capable of generating human-like responses. For this purpose, AI models need extensive training, which requires a large amount of data. Alongside that, OpenAI’s ChatGPT has a weekly user base of 800 million individuals. While managing such a large number of users and training data, ChatGPT needs advanced computing capacity.
The demand for ChatGPT has significantly increased after the launch of GPT-5. OpenAI is further planning to boost its capability with sophisticated computing systems. Sam Altman, the CEO of OpenAI, revealed last month that the firm is planning to double its computing power following the surging demand for GPT-5.
Developing its own chip for AI workloads will enable OpenAI to reduce dependency on its existing chip provider, Nvidia, and reduce chip supply costs.
Discover the latest tech news and insights, exclusively with HiTechNectar!
Recommended For You:
OpenAI Introduces GPT-4o, a Faster and Free Model for all ChatGPT Users