Internet Status — July 26, 2026: Incidents Detected, No Internet-Wide Outage
The internet is up. Core infrastructure is healthy today, but a handful of named incidents are active — most visibly an OpenAI/ChatGPT disruption that is now in recovery.
**Active incidents**
**OpenAI / ChatGPT** — OpenAI's status page lists an open incident, "Elevated errors affecting ChatGPT conversations," currently in the Monitoring phase. OpenAI states: "Mitigation has been implemented, and we are monitoring the recovery of ChatGPT conversations." Press coverage of the disruption described users worldwide hitting login failures, missing conversations and error messages across the ChatGPT website and mobile app, with OpenAI's API and Codex also affected at the peak. The incident is trending toward recovery rather than worsening. (status.openai.com)
**Twilio** — Three open, carrier-level incidents: SMS delivery failures from the US to a subset of Twilio short codes; delayed SMS delivery receipts to Claro Guatemala; and voice call failures to Reliance Jio subscribers in India. Twilio has identified the cause on the latter two. Impact is narrow and route-specific rather than platform-wide. (status.twilio.com)
**Vercel** — The Vercel Community forums are unavailable for maintenance. Hosting, builds and edge delivery are unaffected. (vercel-status.com)
**Ongoing background: AWS Middle East** — The AWS Middle East (UAE) me-central-1 and Middle East (Bahrain) me-south-1 regions remain degraded following physical damage to facilities earlier this year. AWS has said recovery is expected to take months and advises affected customers to run workloads in alternate regions. This is a long-running situation, not a new event today, but it continues to affect customers anchored to those two regions.
**All clear**
Google Cloud reports no broad severe incidents. Microsoft Azure shows no active events, with all regions green. Cloudflare, Akamai and Fastly report normal edge delivery. GitHub, Discord, Atlassian, Slack, Zoom, Heroku, DigitalOcean and Stripe are all fully operational. Cloudflare's status page lists its usual scattering of single-datacenter partial outages and scheduled maintenance windows — routine traffic re-routing that is part of normal operation, not a service incident.
**Recently resolved (past 72 hours)**
Akamai resolved edge delivery issues on July 24. DigitalOcean resolved a network connectivity problem on the India-to-NYC path on July 24. GitHub cleared its open incidents on July 25. An Azure West US connectivity and latency incident on July 23 ran from 14:44 to 19:41 UTC and is closed.
**Assessment**
Status: incidents-detected. The issues in play are single-provider and, in Twilio's case, route-specific. There is no evidence of BGP instability, DNS failure, submarine cable damage or multi-provider correlation. Routing and naming infrastructure look normal.
One caveat on completeness: Downdetector and Fastly's status page both blocked automated requests this run, and Meta's status page did not return readable content. Fastly was confirmed clear via secondary sources; the crowd-sourced signal for this run is thinner than usual.