OpenAI plans to launch its own AI chip

Since at least last year, OpenAI has discussed various options to address the shortage of expensive AI chips it relies on, according to people familiar with the matter.

These options include building its own AI chips, working more closely with other chipmakers including Nvidia, and also diversifying suppliers beyond Nvidia.

OpenAI CEO Sam Altman has made acquiring more AI chips a top priority for the company. He has publicly complained about the shortage of graphics processing units, a market dominated by Nvidia, which controls more than 80% of the global market for chips best suited to running AI applications.

The push to acquire more chips is tied to two major concerns Altman identified: a shortage of advanced processors that power OpenAI's software and the enormous costs associated with running the hardware needed to power AI efforts and products.

Since 2020, OpenAI has been developing general artificial intelligence technologies on a giant supercomputer built by Microsoft, one of its biggest backers, using 10,000 Nvidia graphics processing units (GPUs).

Running ChatGPT is expensive for the company. Each query costs about 4 cents, according to Bernstein analyst Stacy Rasgon.

If ChatGPT queries grew to one-tenth the size of Google searches, it would initially require about $48.1 billion worth of GPUs and about $16 billion worth of chips per year to keep running.

The era of custom chips

The effort to develop its own AI chips would put OpenAI in a small group of large tech companies like Alphabet's Google and Amazon.com Inc. that are looking to gain control over the design of chips for their businesses.

It’s unclear whether OpenAI will move forward with its plans to build custom chips. Doing so would be a major strategic initiative and a major investment that could cost hundreds of millions of dollars a year, according to industry veterans.

Even if OpenAI commits resources to the task, it doesn’t guarantee success. Acquiring a chip company could accelerate OpenAI’s process of building its own chips — much like Amazon.com did with its 2015 acquisition of Annapurna Labs.

OpenAI has been looking closely at that option and has begun due diligence on acquiring a chip company, according to people familiar with the company’s plans. The identity of the company OpenAI is considering acquiring is not yet known.

Even if OpenAI moves forward with its plans for custom chips — including acquisitions — the effort will likely take several years, leaving the company dependent on commercial suppliers like Nvidia and Advanced Micro Devices in the meantime.

Several major tech companies have been building their own processors for years, but results have been limited.

Meta’s efforts to produce custom chips have run into problems, forcing the company to scrap some of its AI chips, Reuters reports. The Facebook owner is now working on a newer chip that can support all types of AI tasks.

OpenAI's main backer, Microsoft, is also developing a custom AI chip that OpenAI is testing.

Demand for dedicated AI chips has skyrocketed since ChatGPT launched last year. AI chips, or accelerators, are needed to train and run the latest generation of AI technology. Nvidia is one of the few chipmakers that can produce useful AI chips and dominate the market.

With this move, Open AI has raised questions about autonomy and control in the development of artificial intelligence technology.

This represents a trend in the AI ​​industry, as many organizations seek to build their own hardware infrastructure to improve performance and control in the development and deployment of artificial intelligence applications.