OpenAI, the company behind ChatGPT, is exploring solutions to address the chip shortages required to drive its artificial intelligence (AI) technology. Among the options currently under consideration, as outlined in a recent media report, is the potential development of its own high-performance processor.
Reportedly, OpenAI is investigating the possibility of creating its own in-house AI chips to address the shortage of GPUs.
OpenAI, the AI research lab, is considering the development of its own processors or AI accelerators to address the ongoing shortage of the expensive hardware required for its operations. This shortage has been exacerbated by the high demand resulting from the launch of its AI-based chatbot last year.
To tackle this issue, several options are being explored, including the acquisition of a chip company and closer collaboration with chip manufacturers, not limited to market leader Nvidia but also involving other suppliers, according to sources with knowledge of the matter.
OpenAI’s CEO, Sam Altman, who has previously expressed concerns about GPU scarcity and associated costs, has made acquiring more AI chips a top priority for the organization.
Even if OpenAI proceeds with its plan to develop its own chip, which has not been finalized yet, the organization is expected to remain reliant on other providers like Nvidia for several years. Currently, Nvidia controls 80% of the supply of processors for AI applications.
OpenAI’s generative AI technology relies on a supercomputer that uses 10,000 of Nvidia’s GPUs, which was built with significant support from Microsoft, one of its major backers. Running the AI assistant is costly, with each query estimated to cost around $0.04.
Should OpenAI move forward with developing its own AI chips, it would join a select group of tech giants that have embarked on similar endeavors, including Google and Amazon, which aim to take control of chip design integral to their businesses.
Meta, the parent company of Facebook, has also explored building its own processor for its AI initiatives, although it has faced challenges in the development process and is now working on a newer model.
Additionally, Microsoft, a major supporter of OpenAI’s chatbot project, is reportedly working on a custom AI chip that is undergoing testing with the ChatGPT maker, as reported by the tech-focused publication The Information.