Just after midnight on Sunday, the feeds stopped. Facebook users trying to refresh saw error screens; Instagram opened to blank timelines. Within minutes, complaints were flooding DownDetector, and on X, users traded workarounds and jokes while waiting for the platforms to come back. Meta restored service after about two hours, according to the company, which said little about what had gone wrong.
Meta acknowledged the disruption in brief statements and said the platforms were back online. It has not explained the cause. Industry analysts pointed to an internal network configuration change, a class of error with a well-documented history at Meta. The company’s most famous outage, in October 2021, took Facebook, Instagram, and WhatsApp offline for about six hours after a routine maintenance change to its network configuration cascaded through its data centers. Analysts at the time estimated the episode cost the company roughly $100 million in revenue.
This time the failure was shorter and narrower. Facebook and Instagram were affected; WhatsApp appeared to keep working. But the timing drew attention. The interruption came as restrictions on TikTok in the United States have eased and as competition among social platforms has intensified. Analysts said even a two-hour outage carries a cost beyond lost ad impressions, because advertisers and users remember downtime when contracts come up for renewal.
The scale of what went dark is hard to grasp. Meta runs one of the largest fleets of data centers in the world, and Facebook and Instagram together serve billions of people. When a configuration change goes wrong at that scale, the failure is not a single server problem but a network-wide event, and recovery means walking the change back through every region in sequence. The speed of this recovery, roughly two hours from first reports to restored service, suggests the company’s response systems worked even if its change controls did not.
For small businesses, the interruption was direct. Many merchants run customer service and order flow through Instagram’s messaging tools, and a multi-hour gap means unanswered inquiries and delayed orders. Advertisers face a subtler problem: an outage distorts campaign delivery for the affected window, and agencies typically seek credits or make-goods afterward, adding friction to a relationship Meta is trying to keep simple.
Regulators are watching. Under the European Union’s Digital Services Act, major platforms face transparency requirements, and Brussels has made clear that severe disruptions to services used by millions of Europeans are the kind of event it intends to monitor. The 2021 outage prompted European officials to press Meta for explanations, and this episode is likely to draw similar questions, analysts said.
The outage was the worst of 2026 for Meta, a year that has otherwise been quiet on infrastructure. The company has spent heavily on reliability, and its internal post-mortems after 2021 led to changes in how it tests and rolls out network configurations. That those safeguards did not prevent a two-hour failure on Sunday shows the limits of even the most rigorous change management when the network is this large.
The competitive context sharpens the stakes. TikTok’s position in the United States has stabilized after years of legal and political turmoil, and Meta has responded by pushing its own short-video products. The outage came the same day the Justice Department lifted the federal ban on TikTok for government employees, an easing of restrictions that had been seen as a headwind for Meta’s rivals. The timing handed competitors an opening, analysts said, even if the outage itself was an accident of engineering rather than a strategic event.
Users who could not refresh their feeds turned elsewhere, and the outage became a trending topic on X within minutes. For Meta, that is the humbling part of the modern platform business: when the service goes down, the conversation about it happens on a competitor’s service, and there is nothing the company can do about it.
The outage also rippled beyond Meta’s own apps. Many third-party services offer sign-in through Facebook or Instagram, and users reported being unable to log in to those services during the window, a sign of how deeply Meta’s identity systems run through the wider internet. Companies that rely on Facebook login for their own products watched the episode with particular unease, analysts said, since a Meta failure can lock their users out too.
Meta said service had returned to normal by the time its brief statements were issued, and most users found their feeds working again within the two-hour window. For a company that tells advertisers it can be trusted with their budgets, two silent hours are a problem no press release can fully fix. The industry’s attention has shifted to the underlying question: whether the networks that carry daily life for billions of people can be made reliable enough that outages become the exception rather than the recurring event they have been.
