OpenAI is Exploring Making Its own AI chips
OpenAI is exploring making its own AI chip, and has gone as far as evaluating a potential target for acquisition, according to a Reuters report.
While it has not decided if it would proceed, the creator of ChatGPT is reportedly mulling this as one of three options. The other options would be to work more closely with chipmakers such as Nvidia, and diversifying suppliers beyond Nvidia.
Nvidia currently controls more than 80% of the chips suited for powering AI applications.
As the world rushes into AI and companies scrambles to secure GPUs to power their AI initiatives, the demand for powerful AI chips is skyrocketing. This has resulted in a global squeeze, even among large technology firms normally immune to such supply chain issues.
Roll your own AI chip
CEO Sam Altman has publicly complained about the scarcity of GPUs , and has made the acquisition of more AI chips a top priority for OpenAI.
Considering OpenAI’s pole position in generative AI today, it is conceivably that the inability to access enough GPUs could cause it to fall behind its competitors.
Of course, even a decision today would likely take years to bear fruition, leaving OpenAI dependent on Nvidia in the meantime.
Going ahead with developing its own AI chip would put OpenAI among a very small group of tech players with the ability to design their own processor. Only the likes of Google and Amazon have sought to design the processors
According to reports, Microsoft is also developing a custom AI chip to reduce its dependence on Nvidia’s GPUs. According to a report published by The Information, Microsoft will likely unveil its first AI chip at its Ignite 2023 conference next month.
Citing sources in the know, the report noted that this “could help Microsoft lessen its reliance on Nvidia-designed AI chips, which have been in short supply as demand for them has boomed.”
Microsoft’s chips will reportedly be designed for use within data centres to train and run large language models, though it isn’t clear if it would be a full-fledged GPU or specialist chip designed more for AI inference.
The development of custom AI chips by major technology companies highlights the growing importance of artificial intelligence in various industries. For OpenAI, creating its own AI chip could potentially provide a competitive edge and ensure a steady supply of hardware to support its ongoing research and development efforts.
Paul Mah is the editor of DSAITrends. A former system administrator, programmer, and IT lecturer, he enjoys writing both code and prose. You can reach him at [email protected].
Image credit: iStockphoto/Jian Fan
