Shares of Meta popped 10% on Wednesday following news that the company is building out a new cloud business that could help recoup some of the billions of dollars it’s poured into artificial intelligence infrastructure.
Meta will sell its excess computing power to outside customers, CNBC’s Jim Cramer confirmed. Bloomberg was first to report the news.
The company is debating whether it will offer access to AI models that are hosted on its infrastructure, or whether it will sell access to raw computing power, according to Bloomberg.
A representative for Meta did not immediately respond to CNBC’s request for comment.
Model developers, including Meta, have been racing to secure computing power since OpenAI kickstarted the AI boom with the launch of its ChatGPT chatbot in 2022, and demand far outpaces supply. Meta told investors in April that it plans to spend as much as $145 billion on capex this year as it continues developing data centers and securing the graphics processing units needed to train AI models and run large workloads.
By standing up a cloud business, Meta could generate revenue on the capacity it’s not using, a welcome signal for some investors who have been uneasy about the company’s spending plans. The new business would also throw Meta into a new and fiercely competitive market, which is dominated by companies including Amazon, Microsoft, Google and CoreWeave, among others.
Shares of neocloud companies CoreWeave and Nebius Group both plunged following the Meta news, sinking about 12% each.
Mark Zuckerberg first signaled the possibility of a cloud move in the company’s Q3 2025 earnings, and addressed it again in May during Meta’s annual shareholder meeting.
“It’s definitely on the table,” Zuckerberg told investors, adding that if Meta gets to a point where it has overbuilt AI infrastructure, “then that is an option that we have.”
Meta is following the lead of Elon Musk’s SpaceX, which has also started selling excess computing capacity this year. The company has inked lucrative deals with Anthropic, which has agreed to pay $1.25 billion per month for capacity, and Google, which has agreed to pay $920 million a month.
Meta has been struggling to find its footing in the AI industry, even after spending $14 billion to bring in Alexandr Wang from Scale AI last year. The company debuted its first model under Wang’s leadership, Muse Spark, in April, which it positioned as a “powerful foundation,” not a state-of-the-art offiering.
—CNBC’s Julia Boorstin contributed to this story.
WATCH: Meta’s Alex Himel on new Meta glasses: Wanted a more accessible price point
