What Caused the June 22, 2026 Cloudflare Outage and Which Services Were Affected

On June 22, 2026, a significant portion of the internet went down. The culprit was not a software bug or a cyberattack — it was a severed fiber line in Eastern North America. Cloudflare, which handles roughly 20% of global web traffic, saw a cascade of failures that took down X, Reddit, Microsoft Teams, Zoom, Amazon Web Services, and dozens of other platforms. Here is the detailed breakdown of what happened, how it compares to previous outages, and what role — if any — Cloudflare's scheduled Newark maintenance played.

A Fiber Cut, Not a Software Bug

Cloudflare engineers first detected elevated error rates and latency at approximately 13:35 UTC on June 22, 2026. By 14:37 UTC, they had traced the root cause to a fiber cut in Eastern North America — a physical break in the underground or undersea cables that carry internet traffic . As of the latest available reports, the fix was still in progress .

This was not an internal configuration mistake, a DDoS attack, or a malicious act. It was an external physical infrastructure failure, making it the first major Cloudflare outage of its kind in this recent series of disruptions .

Which Services Were Affected

Because Cloudflare sits between users and a vast number of websites, a single fiber cut created a domino effect. DownDetector recorded massive spikes in user-reported outages for:

• X — more than 35,000 user reports at peak
• Microsoft Teams, Reddit, Zoom, Robinhood, Canva
• Amazon Web Services, Fortnite, Discord

Cloudflare's own status page showed a "Partial System Outage" with 54 of 457 components experiencing disruption as of late on June 22 . The company acknowledged "increased error rates and latency" across multiple services .

NetBlocks, which monitors global internet health, noted that X was "experiencing international outages for a second consecutive day amid timeline backend API errors," but clarified the incident was "not related to country-level internet disruptions" .

Was There a Separate X Outage on June 21?

No. The available sources consistently describe X going down on June 22 as part of the same Cloudflare-linked event . There is no corroborated evidence of a significant X-specific outage on June 21, 2026 that was separate from this incident. If you saw reports of an X outage on June 21, that is not supported by the current reporting .

The Role of Cloudflare's Scheduled Newark Maintenance

Cloudflare had scheduled maintenance in its EWR (Newark) datacenter running from June 22, 2026, 01:00 UTC through June 23, 2026, 09:00 UTC . The status page noted that traffic might be re-routed from Newark during this window, potentially causing a "slight increase in latency" for end-users in the affected region .

However, the scheduled maintenance was not identified as the cause of the fiber-cut outage. The two incidents overlapped in timing but were separate in root cause — the Newark maintenance was routine and planned, while the fiber cut was an unforeseen external infrastructure failure that struck later in the day .

Broader Pattern: A Year of Cloudflare Outages

The June 22 fiber-cut incident is the latest in a string of notable Cloudflare disruptions since late 2025. Here is a comparison:

Date	Duration	Cause	Scale
November 18, 2025	~4 hours	Oversized Bot Management configuration file triggered a cascading failure (initially mistaken for a DDoS attack)	Roughly one in five webpages affected
December 5, 2025	~25 minutes	Internal network failure	~28% of all Cloudflare HTTP traffic
February 22, 2026	~6 hours	Internal BGP configuration update unintentionally withdrew customer routes	25% of all BYOIP prefixes globally
June 22, 2026	Ongoing (as of last report)	Fiber cut in Eastern North America	54 out of 457 components disrupted; multiple major platforms affected

The common thread: Cloudflare's position as infrastructure for roughly 20% of global web traffic means that both external physical damage (fiber cuts) and internal configuration errors (oversized config files, BGP mistakes) can produce cascading failures across hundreds of major platforms. The June 22 fiber-cut outage is the first major physical-infrastructure cause in this recent series, as opposed to software or configuration bugs.

Why This Matters

A single fiber cut in one region was enough to disrupt services that tens of millions of people use every day. This incident underscores a fundamental vulnerability in the internet's architecture: even the most resilient cloud providers depend on physical cables that can be damaged by construction, weather, or accidents. For businesses and users, it is a reminder that redundancy at the application layer — multi-cloud strategies, failover systems, and offline fallbacks — remains essential even when working with a provider as dominant as Cloudflare.

As of the latest reports, the fix for the fiber cut was still in progress . Cloudflare has not yet published a full post-mortem, and it is not yet clear whether the fiber cut was accidental, weather-related, or caused by construction work.