OpenAI is working on its own chipset, dependence on any other company will end

OpenAI plans to build its first custom artificial intelligence (AI) chipset this year. According to a report, this San Francisco-based AI company has started the process of designing this chipset internally and is moving towards completing its design in the next few months. The main objective of OpenAI in developing this chipset is to reduce its dependence on Nvidia and strengthen its bargaining position with other chip suppliers. A recent trademark filing also revealed that the company is planning to manufacture not only chipsets but many types of hardware.

OpenAI is completing its internal design of the chipset and will complete it within a few months' time. According to the report, once the design is completed, the company will send it to Taiwan Semiconductor Manufacturing Company (TSMC), which will manufacture this chipset.

This latest OpenAI chipset will be a 3-nanometer process technology-based and feature high-bandwidth memory (HBM) with systolic array architecture and advanced networking capabilities. Interestingly, HBM-based designs are also used in Nvidia's chipsets as well.

The company believes that manufacturing its own chipset will give it a strategic edge in bargaining with other chipset suppliers. Also, it will also reduce its excessive dependence on Nvidia, whose chips OpenAI has been using extensively. According to the report, the company plans to develop processors with more advanced and wider capabilities in the future.

According to the report, this chipset of OpenAI is being designed by an in-house team led by Richard Ho, who is the hardware head of OpenAI. It is worth noting that Ho has previously worked at Lightmatter and Google and has expertise in semiconductor engineering. The report also states that Ho's team has doubled in the last few months and now includes 40 employees.