OpenAI, the parent company of ChatGPT, is trying to find new ways to its own artificial intelligence chips and has gone much ahead as calculating a potential takeover target, according to people who are well versed with the company’s plans.
People familiar with the matter say that the company has not yet decided to move ahead, as per recent internal discussions described to Reuters. Moreover, since at least in 2022, it discussed various options to solve the shortage of expensive AI chips that OpenAI depends on.
These options constitute of building its own AI chip, working more closely with other chipmakers including Nvidia and also broadening its spectrum in supplies beyond Nvidia.
OpenAI refused to comment on the matter.
For CEO Sam Altman, the takeover of more AI chips a top priority for the company. He has publicly complained about the limitation of graphics processing units, a market dominated by Nvidia, which heads over 80 per cent of the global market for the chips best suited to run Artificial Intelligence applications.
The need to get more chips is mainly attributed to two important concerns which CEO Altman has recognised: a shortage of the advanced processors that power OpenAI’s software and the “eye-watering” costs linked with running the hardware that are required to elevate its efforts and products.
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 Nvidia’s graphics processing units (GPUs).
Running ChatGPT is causing the company a lot. Every complaint costs approximately 4 cents, as per a report conducted by Bernstein analyst Stacy Rasgon. If ChatGPT complaints grow to a tenth the scale of Google search, it would cost around $48.1 billion worth of GPUs first and almost $16 billion worth of chips every year to keep it working.
An effort to develop its own AI chips would put OpenAI among a small group of big tech behemoths like Alphabet’s Google and Amazon.com that have sought to take control over designing the chips that are paramount to their businesses.
It is not whether OpenAI will move ahead with a plan to build a custom chip. Doing so would be a major strategic initiative and a heavy investment that could amount to hundreds of millions of dollars a year in costs, according to industry veterans. Even if OpenAI committed resources to the task it would not guarantee success.
The identity of the company OpenAI probed in for purchasing could not be known.
Even if OpenAI goes ahead with its goal for a custom chip – including a takeover – the effort is likely to take several years leaving the company dependent on commercial providers like Nvidia and Advanced Micro Devices in the meantime.
According to a report by Reuters, some big tech titans have been creating their own processors for years with no significant results. Meta’s custom chip effort has been detected with issues, which has subsequently led the company to scrap some of its AI chips. The Facebook owner is now creating on a newer chip that will cover all types of AI work.
The Information has reported that OpenAI’s main supporter, Microsoft, is also developing a custom AI chip that OpenAI is testing. The plans also indicate further distancing between the two companies.
A sudden surge in demand for specialized AI chips has occurred ever since the launch of ChatGPT in 2022. Particular chips, or AI accelerators, are needed to train and run the recent generative AI technology. Nvidia is one of the scarce chipmakers that produces useful AI chips and has monopoly over the market.
