OpenAI Plans Additional $100 Billion In Server Spending To Ease Compute Crunch: Report

OpenAI is reportedly preparing to dramatically expand its spending on computing power as demand for its artificial intelligence products continues to strain infrastructure. 

According to a report by The Information, the company informed some of its shareholders that it plans to spend an additional $100 billion on rented backup servers over the next five years. The report added, citing sources, that the new expenditure would be in addition to the $350 billion already projected through 2030 for server rentals. On Stocktwits, retail sentiment around OpenAI, which is not publicly traded, was in ‘bullish’ territory amid ‘high’ levels of chatter on Friday.

The report stated that the additional capacity is intended to alleviate persistent shortages of computing resources that have forced the company to delay product rollouts and throttle certain features. 

OpenAI Chief Financial Officer Sarah Friar highlighted the server capacity challenge last week at a Goldman Sachs conference, reportedly stating that the company is “massively compute constrained.” Friar said OpenAI has repeatedly had to hold back launches of new models or intentionally slow some services because it lacks enough servers to handle surges in usage.

Even last year, OpenAI CEO Sam Altman admitted during a Reddit AMA that a lack of compute capacity is one of the major factors preventing the company from shipping products as often as it would like.

The company has often capped the availability of new products amid server capacity shortfalls. Its March launch of ChatGPT’s photo-to-animation features was the latest example, when overwhelming demand triggered temporary restrictions.

Read also: Tim Cook Reportedly Says Apple Watch Could Alert 1 Million Users to Hypertension This Year

For updates and corrections, email newsroom[at]stocktwits[dot]com.