San Francisco: ChatGPTs developer OpenAI is considering the development of its own artificial intelligence chips and has even explored a potential acquisition target, Reuters reported, citing sources.
While no definitive decision has been reached, recent internal discussions have centered around finding a solution to the scarcity of expensive AI chips that OpenAI depends on, the report said
CEO Sam Altman has prioritized acquiring more AI chips, expressing concerns over the shortage of graphics processing units, a market largely controlled by Nvidia.
Since 2020, OpenAI has developed its generative artificial intelligence technologies on a massive supercomputer constructed by Microsoft, one of its largest backers, that uses 10,000 of Nvidias graphics processing units (GPUs), stated the Reuters report.
Obtaining additional chips is crucial due to the shortage of advanced processors powering OpenAIs software and the substantial costs associated with running the necessary hardware.
OpenAI has been using a massive supercomputer with 10,000 of Nvidias graphics processing units, a setup backed by Microsoft, one of its major backers.
According to the Reuters report, operating ChatGPT is costly, with each query amounting to about 4 cents. If ChatGPTs queries were to reach a tenth of Googles search scale, it would initially require approximately $48.1 billion worth of GPUs and about $16 billion worth of chips annually to sustain operations.
Even if OpenAI proceeds with custom chip plans, including a potential acquisition, the process is expected to take several years, making the company reliant on commercial providers like Nvidia and Advanced Micro Devices in the interim, the report said.
Some major tech companies have attempted to develop their own processors with limited success.
Meta, for example, encountered difficulties with its custom chip initiative, leading to the abandonment of some AI chips. Meta is now working on a newer chip designed to cover various types of AI work.
Microsoft, OpenAIs primary supporter, is also working on a custom AI chip that OpenAI is testing, suggesting a potential divergence between the two companies, according to the report.
The demand for specialized AI chips has multiplied since the introduction of ChatGPT, as specific chips or AI accelerators are essential for training and operating the latest generative AI technology.
Nvidia remains one of the few chipmakers that produces valuable AI chips and dominates the market, the report said.
