- Oracle says severe winter weather knocked out power at one of its U.S. data centres, leading to cascading failures that disrupted TikTok services.
- The outage coincides with TikTok’s shift to U.S. ownership and has reignited concerns over censorship, reliability and infrastructure resilience.
What Happened: Power outage knocks TikTok offline amid U.S. winter storm
A powerful winter storm sweeping parts of the United States over the weekend triggered a temporary power failure at one of Oracle’s data centres, the cloud computing firm confirmed this week. According to Oracle, the outage was the root cause of technical issues experienced by TikTok users, including slowed load times, posting errors and disruptions to the personalised “For You” feed.
The outage reportedly caused a “cascading systems failure” that affected core app functions and other services hosted at the same facility, leaving users with zero views on new uploads and incomplete engagement metrics. TikTok’s new U.S.-based joint venture handling American operations has been working with the data centre provider to restore full service, but as of Tuesday some performance issues remained unresolved.
This disruption struck just days after TikTok finalised a deal creating the TikTok USDS Joint Venture LLC, in which Oracle, private equity firm Silver Lake and MGX each hold a stake alongside majority U.S. and global investors — a move designed to satisfy U.S. regulatory requirements and avert a ban on the platform.
Also Read: TikTok signs long-awaited deal to divest U.S. unit to American investors
Also Read: USA TikTok ban: Empty threat or necessary measure?
Why It’s Important
While Oracle and TikTok maintain the outage was strictly technical, the timing has fuelled speculation online and among some political figures that the issues might be linked to content moderation or platform control, particularly for politically sensitive posts. California’s governor suggested possible “suppression” of content critical of prominent political figures — claims both companies deny.
The incident also raises broader questions about infrastructure resilience under the new U.S.-centric model. Concentrating critical data and algorithm operations in a limited set of facilities may create single points of failure; critics ask whether reliance on a single provider like Oracle was adequately risk-assessed, especially given that other major platforms typically distribute workloads across multiple redundant cloud systems.
Lastly, trust among creators — who depend on predictable engagement and visibility for livelihoods — may be eroded if outages like this persist or recur shortly after ownership transitions, particularly as the platform navigates legal, regulatory and public perception challenges in one of its largest markets.
