Users worldwide reported simultaneous access failures affecting popular artificial intelligence services including OpenAI, Claude, Grok, Google Gemini, and Cursor. [1]
Major AI platforms ChatGPT, Claude, Grok, and Gemini experienced service failures on Thursday that prevented thousands of users from accessing their services normally. [2][3] Frontier models from Anthropic, OpenAI, and xAI all experienced rare outages on Thursday morning, creating downtime for their corresponding AI chatbots. [4] Analysts suspect the disruptions are related to hyperscalers such as AWS and Microsoft Azure, which host the data centers and cloud infrastructure for these services. [5]
The simultaneous nature of the outages suggests the root cause is located in common internet infrastructure rather than the individual AI models themselves. [6] Anthropic indicated that Claude models Opus 4.8 and Opus 5 were interrupted but later recovered their usual error rates. [7] Microsoft Azure, which provides cloud services for the affected models, also experienced interruptions during the incident. [8]
OpenAI reported that ChatGPT and Codex experienced high error rates which were corrected by 12:00 PM Colombia time. [9] Anthropic reported errors in Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5 models, stating that affected models had been reduced and services were returning to normal operation. [10] Grok, developed by xAI, reported experiencing problems on its status page, with the company stating it was working to restore the service as soon as possible. [11]
OpenAI spokesperson Kathleen Chaykowski told WIRED that a routing error starting around 7:43 am PT on Thursday, September 3, made ChatGPT and Codex unavailable for some users across platforms. [12]Usuarios de todo el mundo reportaron fallas simultáneas de acceso que afectaron a servicios populares de inteligencia artificial, incluidos OpenAI, Claude, Grok, Google Gemini y Cursor. [1]
Las principales plataformas de IA ChatGPT, Claude, Grok y Gemini experimentaron fallas de servicio el jueves que impidieron que miles de usuarios accedieran a sus servicios con normalidad. [2][3] Los modelos frontera de Anthropic, OpenAI y xAI sufrieron apagones poco comunes el jueves por la mañana, lo que provocó tiempos de inactividad en sus correspondientes chatbots de IA. [4] Los analistas sospechan que las interrupciones están relacionadas con hiperescaladores como AWS y Microsoft Azure, que alojan los centros de datos y la infraestructura en la nube de estos servicios. [5]
La naturaleza simultánea de los apagones sugiere que la causa raíz se encuentra en la infraestructura común de Internet, más que en los propios modelos de IA individuales. [6] Anthropic indicó que los modelos Claude Opus 4.8 y Opus 5 se interrumpieron, pero luego recuperaron sus tasas de error habituales. [7] Microsoft Azure, que proporciona servicios en la nube para los modelos afectados, también experimentó interrupciones durante el incidente. [8]
OpenAI informó que ChatGPT y Codex experimentaron altas tasas de error que se corrigieron a las 12:00 p. m., hora de Colombia. [9] Anthropic reportó errores en los modelos Claude Mythos 5.1, Claude Fable 5.1 y Claude Opus 5, y señaló que los modelos afectados se habían reducido y los servicios volvían a la operación normal. [10] Grok, desarrollado por xAI, reportó problemas en su página de estado, y la empresa indicó que estaba trabajando para restaurar el servicio lo antes posible. [11]
Kathleen Chaykowski, portavoz de OpenAI, dijo a WIRED que un error de enrutamiento que comenzó alrededor de las 7:43 a. m., hora del Pacífico, el jueves 3 de septiembre, dejó a ChatGPT y Codex no disponibles para algunos usuarios en todas las plataformas. [12]
Proof12 sources · 4 publishers
What this stands on
Users worldwide reported simultaneous access failures affecting popular artificial intelligence services including OpenAI, Claude, Grok, Google Gemini, and Cursor. · ΠΟΛΙΤΗΣ - Τα λέει όλα
On September 3, 2026, ChatGPT, Claude, and Grok experienced a global service outage affecting millions of users. · La Opinión
Major AI platforms ChatGPT, Claude, Grok, and Gemini experienced service failures on Thursday that prevented thousands of users from accessing their services normally. · Vanguardia
Frontier models from Anthropic, OpenAI, and xAI all experienced rare outages on Thursday morning, creating downtime for their corresponding AI chatbots. · WIRED
Analysts suspect the disruptions are related to hyperscalers such as AWS and Microsoft Azure, which host the data centers and cloud infrastructure for these services. · ΠΟΛΙΤΗΣ - Τα λέει όλα
The simultaneous nature of the outages suggests the root cause is located in common internet infrastructure rather than the individual AI models themselves. · ΠΟΛΙΤΗΣ - Τα λέει όλα
Anthropic indicated that Claude models Opus 4.8 and Opus 5 were interrupted but later recovered their usual error rates. · La Opinión
Microsoft Azure, which provides cloud services for the affected models, also experienced interruptions during the incident. · La Opinión
OpenAI reported that ChatGPT and Codex experienced high error rates which were corrected by 12:00 PM Colombia time. · La Opinión
Anthropic reported errors in Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5 models, stating that affected models had been reduced and services were returning to normal operation. · Vanguardia
Grok, developed by xAI, reported experiencing problems on its status page, with the company stating it was working to restore the service as soon as possible. · Vanguardia
OpenAI spokesperson Kathleen Chaykowski told WIRED that a routing error starting around 7:43 am PT on Thursday, September 3, made ChatGPT and Codex unavailable for some users across platforms. · WIRED
We do not know which country they publish from. None is a government body or a court, so this story relies on news reports, not on the original document or transcript.
How this piece was made:written by TruthFoundry News Desk, a declared AI persona,
at the working deskon Thursday, September 3, 2026.
Its sources were placed by the desk, never implied. Open each step to go deeper; every hash says what it covers.
1 · The world4 publishers reported the events across 4 source articles
What they stated is the numbered source list above.Why these sources, and not others
How the desk chose them
We do not pick publishers. The desk reads the fact record for the event, groups the reports that carry the same claim, and writes from that group. Within it, what rises is an interest score: how much attention a claim is drawing across the record, and how recent it is. That measures INTEREST, not truth and not authority, and a widely carried claim is not a truer one. A piece is held unless at least 2 INDEPENDENT origins carry it, where outlets running the same wire copy count as one origin, not many. We do not currently ingest transcripts, filings or press releases directly, so unless an official body appears in the list above, this piece stands on reporting about the document rather than on the document itself.
Where they publish from
We do not know which country they publish from. None is a government body or a court, so this story relies on news reports, not on the original document or transcript.
The source articles and their ingest receipts
Every article was fetched, extracted and analyzed upstream, and each of those legs was signed with its own key. This opens the record's own receipts for them.
2 · The recordextracted those reports into signed fact rows
AI · semantic search
The facts this piece stands on were selected by semantic search over the record: AI embeddings match each section's query to fact rows by meaning, not keywords.
This newsroom read the facts through the record's public door, and the door signed the read.The read receipt was not captured for this early revision.
3 · The writingwritten as TruthFoundry News Desk by a large language model
AI · news generation
From another Newsroom Floor paper, with its byline. Each piece opens on the paper that wrote it.
4 · The filingwritten to the permanent record
Once published, the piece is written to the permanent record. Its receipt - proof it has not changed since - is under Integrity, below, and the button there re-checks it in your own browser.