Exclusive: ChatGPT-owner OpenAI is exploring making its own AI chips

SAN FRANCISCO/WASHINGTON, Oct 5 (Reuters) – OpenAI, the company behind ChatGPIT, is exploring making its own artificial intelligence chips and has reached out to evaluate a potential acquisition target, according to people familiar with the company’s plans. .

The company has not yet decided how to proceed, according to recent internal discussions seen by Reuters. However, for at least last year it has discussed various options to solve the shortage of expensive AI chips on which OpenAI depends, according to people familiar with the matter.

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

OpenAI declined to comment.

CEO Sam Altman has made the acquisition of 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 effort to acquire more chips is tied to two key concerns identified by Altman: the shortage of advanced processors powering OpenAI’s software and the “eye-watering” associated with running the hardware needed to power its efforts and products. Wally” cost.

Since 2020, OpenAI has developed its generative artificial intelligence technologies on a giant supercomputer built by Microsoft, one of its biggest supporters, which uses 10,000 graphics processing units (GPUs) from Nvidia.

Running ChatGPT is very expensive for the company. Each query costs about 4 cents, according to an analysis by Bernstein analyst Stacey Rasgon. If ChatGPT queries grew to one-tenth the scale of Google Search, it would require about $48.1 billion worth of GPUs initially and about $16 billion worth of chips per year to keep it running.

custom chips era

A keyboard is placed in front of the OpenAI logo displayed in this illustration taken on February 21, 2023. Reuters/Dado Ruvik/Illustration/File Photo Get licensing rights

The effort to develop its own AI chips would put OpenAI among a small group of big tech players like Alphabet’s Google (GOOGL.O) and Amazon.com (AMZN.O) that have sought to take control of designing basic chips. Has demanded. For their businesses.

It’s unclear whether OpenAI will move forward with plans to make custom chips. According to industry leaders, doing so would be a major strategic initiative and a huge investment that could cost millions of dollars per year. Even if OpenAI dedicated resources to the task, it still would not guarantee success.

The acquisition of a chip company could accelerate OpenAI’s process of making its own chips – as happened for Amazon.com and its acquisition of Annapurna Labs in 2015.

According to one of the people familiar with its plans, OpenAI had considered its way to the point where it had conducted due diligence on a potential acquisition target.

The identity of the company purchased by OpenAI could not be known.

Even if OpenAI moves forward with plans for a custom chip — including an acquisition — the effort is likely to take several years, leaving the company reliant on commercial providers like Nvidia and Advanced Micro Devices (AMD.O) in the meantime. .

Some big tech companies have been building their own processors for years with limited results. Meta’s (META.O) custom chip effort has been hit by problems, leading the company to scrap some of its AI chips, according to a Reuters report. The Facebook owner is now working on a new chip that will cover all types of AI functions.

The information noted that OpenAI’s main supporter, Microsoft (MSFT.O), is also developing a custom AI chip that OpenAI is testing. These plans may indicate further distances between the two companies.

Demand for specialized AI chips has increased since the launch of ChatGPIT last year. Specialized chips, or AI accelerators, are needed to train and run the latest generative AI technology. Nvidia is one of the few chip makers that produces useful AI chips and dominates the market.

Our Standards: The Thomson Reuters Trust Principles.