Claude outage in U.S.: login errors and degraded performance reported before fix deployed

August 17, 2026

What failed and when

Claude users encountered authentication failures and slow or incomplete responses on Sunday, August 16. BleepingComputer, citing Anthropic’s status page, reported the first alert at 21:58 UTC, when the company said it was investigating sign‑in problems affecting Claude.ai, Claude Code, and Claude Cowork. Minutes later, at 22:07 UTC, Anthropic logged a broader service disruption, noting degraded performance on Claude.ai and platform.claude.com while the investigation continued. At that point, BleepingComputer said the status page still showed Claude Console and the Claude API as operational.

Unite.AI presented a slightly different snapshot, reporting that the first notice was posted at 21:58 UTC and that the incident was marked resolved at 22:34 UTC, implying roughly 36 minutes of disruption. In its account, Unite.AI listed five affected products—claude.ai, Claude Console, the Claude API, Claude Code, and Claude Cowork—reflecting a wider scope than the early view described by BleepingComputer.

Who was affected and at what scale

Asbury Park Press characterized the impact as widespread, reporting more than 11,000 user complaints shortly after 6 p.m. Eastern and that the outage was trending on Google. The paper added that the app drew the largest share of complaints, with issues also noted on the website and in Claude Chat.

Hindustan Times offered a higher snapshot minutes later: by about 6:15 p.m. EDT, the outlet reported more than 15,000 user complaints. The same report described users across the United States encountering an “authentication service unavailable” message when attempting to access the chatbot. BleepingComputer’s write‑up of user experience aligned with those accounts, citing difficulty signing in, pages failing to load, and requests that did not complete.

These tallies are point‑in‑time readings from third‑party complaint aggregators and newsroom checks, and the precise number of affected users is not established in the sources. The geographic framing—“across the United States”—is how Hindustan Times summarized the pattern of reports; no publisher provided a formal footprint beyond that description.

What Anthropic said and the restoration sequence

According to Hindustan Times’ citations of Anthropic’s status page, the operator first acknowledged the authentication issue, then said “a fix has been deployed” and that it was monitoring recovery before posting that the incident “has been resolved.” That progression matches the timestamps described by BleepingComputer—investigation at 21:58 UTC and a degraded‑performance update at 22:07 UTC—and the closure time reported by Unite.AI at 22:34 UTC.

There is a notable difference between sources on which products were impacted. BleepingComputer’s early snapshot showed Claude Console and the Claude API as still operational while Claude.ai, Claude Code, and Claude Cowork were in a major‑outage state. By contrast, Hindustan Times and Unite.AI both included the API and Console in the list of affected services when summarizing the resolution post. The sources do not reconcile that discrepancy, suggesting that impact visibility evolved over the course of the incident or that the outlets captured the status page at different moments.

Cause and what remains unclear

None of the outlets reported a root cause. Both Hindustan Times and BleepingComputer stated that Anthropic had not disclosed why the outage occurred at the time they published. Unite.AI focused on timing and scope rather than origin. No publisher cited a detailed post‑incident analysis.

Restoration is described in the sources as the operator declaring the issue resolved; none of the reports claim a verified, user‑level confirmation beyond the status page notices they reference. Figures for affected users and the list of impacted products are snapshots that varied between reports and may not represent a final tally.

Monitoring services you depend on

For a wider view of developing disruptions, see the Internet Outages archive and the website and service directory. Status Is Down plans help teams watch the services and websites they rely on, and readers can sign up to choose the coverage that fits.

Sources

Image credit: Image credit: Rafael Henrique | Dreamstime.com

Find a service or website