OpenAI’s Plan to Develop AI Hardware

MetaversePlanet March 18, 2024

0 1 minute read

OpenAI, a key player in the artificial intelligence arena, is gearing up to make strides in developing its own AI hardware.

The organization’s objective is to create tailored solutions without having to rely on external hardware providers like Nvidia. This initiative marks OpenAI’s effort to gain independence in the hardware sector.

OpenAI has started talks to produce its own processors

According to Tom’s Hardware, OpenAI is engaging in discussions to manufacture its own artificial intelligence processors, which are expected to be Neural Network Processing Units (NPUs). These NPUs, dedicated to AI tasks, resemble how GPUs offload graphics processing from the CPU, similarly freeing up the CPU from AI-specific computations by handling them more efficiently.

Should OpenAI’s venture into creating NPUs succeed, the organization might phase out its reliance on Nvidia’s hardware in favor of its custom-designed NPUs. This shift could lead OpenAI to eventually offer its hardware for sale or lease to other companies for public application. While it’s unlikely that OpenAI-branded hardware will hit consumer retail outlets, the possibility remains open.

Nvidia, a frontrunner in the AI hardware sector and the third most valuable company in the US, has been the go-to choice for entities venturing into AI, including OpenAI and Microsoft’s AI projects, which have utilized Nvidia’s hardware. Yet, the trend towards developing in-house solutions signals a desire to reduce dependency on Nvidia’s offerings.

Nvidia, however, is not standing still amidst these developments. The company has introduced Chat With RTX, an AI chatbot that leverages the GPUs in existing systems, positioning itself as a solutions provider even as it faces potential shifts in the AI hardware market. This move indicates that Nvidia remains a contender in the AI space, even as companies like OpenAI seek to develop their own hardware solutions.