Meta plans to bring around 1GW of compute online this year, and expects to field a fleet of more than 1.3 million GPUs by year-end.

CEO Mark Zuckerberg said that the company plans to invest $60-65 billion on capital expenditure, primarily on data centers and servers, as well as grow its AI teams "significantly."

Last year, Meta's capex was between $35bn and $40bn, representing a significant increase. This month, Microsoft said that it would spend $80bn on AI data centers.

In a Facebook post, Zuckerberg also re-announced a $10 billion data center in Richland Parish, northeast Louisiana.

"Meta is building a 2GW+ data center that is so large it would cover a significant part of Manhattan," he said.

When it was first announced in December, the company said that the "custom-designed four million-square-foot campus will be our largest data center to date."

Early site work has begun, with construction set to continue through 2030.

"This will be a defining year for AI," Zuckerberg said. "This is a massive effort, and over the coming years, it will drive our core products and business, unlock historic innovation, and extend American technology leadership."

He also added that the company has "the capital to continue investing in the years ahead," a possible reference to criticism leveled at OpenAI's $500bn Stargate Project, which has yet to prove it will actually be able to raise funds.

Similarly, there may be a deeper meaning in the scale and image used by Zuckerberg: City documents seen by Bloomberg show that the first Stargate project will be the size of Central Park, itself just a section of Manhattan.