TECH NEWS – OpenAI, the entity behind ChatGPT, plans to develop its own artificial intelligence chip, and also intends to acquire additional chip manufacturing companies to address the shortage. Discussions regarding the possible solutions to the costly AI chips shortage, which OpenAI is dependent on due to the industry-wide chip scarcity, have been ongoing since last year.

 

The company considered several options, including:

– Manufacturing its own AI chip,

– Closer collaboration with other chip manufacturers, like Nvidia, and

– Involvement of other suppliers.

Nvidia largely dominates the sector, with the tech company holding an 80% market share in chips capable of running AI applications.

 

According to CEO Sam Altman, the acquisition of further chips is driven by two main factors: the lack of advanced processors required to operate OpenAI’s software, and the high costs associated with operating the hardware necessary for their products. Since 2020, OpenAI has been utilizing a massive supercomputer built by Microsoft, employing Nvidia’s 10,000 GPUs for the development of generative artificial intelligence technologies.

Maintaining ChatGPT operations incurs significant costs for OpenAI. Bernstein analyst Stacy Rasgon estimates that each query costs about 4 cents. Should ChatGPT queries reach one-tenth of Google searches, a substantial initial investment in GPUs would be required, along with significant annual costs related to chips.

If OpenAI decides to develop its own AI chips, it would join tech giants like Google and Amazon, who have aimed for control over the critical business chips market. However, developing a unique chip would be a significant strategic move, requiring potentially hundreds of millions of dollars in investment annually, with success not being guaranteed. Acquiring a chip manufacturing company could expedite the process, as Amazon did with the purchase of Annapurna Labs in 2015.

Source: Reuters