ChatGPT-owner OpenAI is exploring making its own AI chips
OpenAI, the company behind ChatGPT, is considering developing its own artificial intelligence platform and is even considering an acquisition, according to people familiar with the company’s plans.
The company has not yet decided whether to continue, according to recent internal discussions reported by Reuters. However, at least last year, it discussed various options to solve the dearth of AI that OpenAI relies on, according to people familiar with the matter.
These options include building its own AI chip, working closely with other manufacturers including Nvidia, and expanding its suppliers beyond Nvidia. OpenAI declined to comment.
CEO Sam Altman has made acquiring more AI chips a top priority for the company. He has publicly complained about the scarcity of the graphics processing unit, a market dominated by Nvidia, which controls more than 80% of the world market for the best chips for AI applications.
The effort to get more chips is connected to two main factors that Altman discovered: the lack of high-quality developers that make the OpenAI software cost “too much” combined with the use of hardware necessary to make his efforts and its products are powerful.
Since 2020, OpenAI has released its artificial intelligence technology on a large supercomputer built by Microsoft, one of its sponsors, which uses 10,000 graphics processing units (GPUs) from Nvidia.
Running ChatGPT costs the company a lot of money. Each question costs about four cents, according to Bernstein analyst Stacy Rasgon. If ChatGPT queries grow to one-tenth the size of Google Search, it will initially require up to $48.1 billion in GPUs and up to $16 billion per year to keep it running.
TIME CHIPS CULTURE
The effort to develop its own page AI will place OpenAI among smaller technology players such as Google Alphabet and Amazon.com who want to control the design of the main page for their businesses. It is unclear whether OpenAI will continue with the project to build a custom chip.
That would be a major strategic and capital investment that could cost hundreds of millions of dollars annually, according to industry veterans. Even if OpenAI commits resources to this project, it cannot guarantee success.
Acquiring a developer can speed up the process of creating an OpenAI chip – as was the case with Amazon.com and the acquisition of Annapurna Labs in 2015.
OpenAI is evaluating that approach to the extent that it works well with acquisition targets, according to one of the people familiar with its plans.
The identity of the company examined by OpenAI cannot be identified at the time of purchase.
Even if OpenAI continues with its traditional chip plans — including acquisitions — the effort will take years, leaving the company dependent on commercial suppliers like Nvidia and Advanced Micro Devices for now. a.
Some big tech companies have been working on their own for years, with little success. Meta’s culture-breaking efforts have faced problems, leading the company to abandon some of its AI chips, according to a Reuters report.
The owner of Facebook is currently working on a new chip that will cover all kinds of AI services. OpenAI‘s sponsor, Microsoft, is also developing a custom AI chip that OpenAI is testing, The News reported. These plans can show a great distance between the two companies.
Demand for unique AI pages has increased since ChatGPT launched last year. A specific set of papers, or AI tools, is necessary to train and implement innovative AI technologies. Nvidia is one of the few chipmakers that produces valuable AI chips and dominates the market.