TikTok U.S. seeks four gigawatts of computing capacity by early 2030s, CNBC reports

TikTok’s U.S. operation is seeking computing capacity for the coming years, with one source describing a target of four gigawatts by the early 2030s. CNBC reported the search on October 23, citing two people familiar with the discussions. The company says Oracle will remain its primary U.S. cloud infrastructure provider.

The business, formally called TikTok USDS Joint Venture, has sounded out opportunities to obtain multiple gigawatts of capacity. Both people spoke anonymously because they were not authorized to share the information.

The figure describes a reported goal for future capacity, not computing resources already secured. TikTok’s public response addressed its cloud relationship rather than confirming the size or timing of the search.

runs primarily on Oracle Cloud Infrastructure in the U.S. and Oracle will continue to be our primary cloud infrastructure provider.

TikTok U.S. spokesperson, speaking to CNBC

Oracle’s role goes beyond cloud hosting

Oracle is also one of the venture’s three managing investors, alongside Silver Lake and MGX. Each holds 15%. TikTok USDS Joint Venture was established in 2026 to comply with President Donald Trump’s executive order requiring the U.S. business to be divested from Chinese parent ByteDance, which retained a 19.9% stake.

The connection also covers TikTok’s recommendation algorithm. In January, the U.S. venture said it would retrain, test and update that system using U.S. user data. It said the algorithm would be secured in Oracle’s U.S. cloud environment. That earlier plan gives the reported capacity search a concrete operational backdrop: the venture already has machine-learning work tied to Oracle.

AI hiring offers clues, not a capacity breakdown

CNBC also pointed to numerous machine-learning vacancies on TikTok U.S.’s website as signs of work that could require more computing capacity. One job advertisement describes the company as shipping a flagship business-to-business AI product focused on running large models and automating creative work.