The three major cloud companies collapsed at the same time? Cursor, ChatGPT followed suit, netizens: The entire Internet is going to be ruined

Last night Beijing time, on X, netizens claimed that AWS, Google Cloud, Azure, and Cloudflare experienced an outage simultaneously.

The outage tracking website Down Detector showed that around 11:30 AM Pacific Summer Time, Google Cloud reported over 13,000 incidents.

As of 11:49 AM Pacific Summer Time on Thursday, cloud computing giant Microsoft Azure recorded approximately 1,000 interruption reports on Down Detector, which dropped to 251 by 12:49 PM Pacific Time.

Around the same time, AWS recorded about 5,000 interruption reports on Down Detector.

However, Microsoft and AWS status pages showed no issues.

According to Down for Everyone, Character AI received around 4,000 interruption reports by 11:19 AM Pacific Summer Time on Thursday.

Additionally, OpenAI, the creator of ChatGPT, stated that they encountered some issues related to single sign-on "and other login methods". The company said in a social media post that their "engineering team is working to mitigate these issues". Moreover, users reported numerous errors with Claude Sonnet 4 (Cursor) and Gemini Pro.

Google Cloud Services Globally Paralyzed for 3 Hours

Among all interrupted services, Google Cloud Services experienced the most severe global outage, causing a series of interruptions for Google's own and third-party applications and platforms.

At the time, the outage tracking website Down Detector's peak began to surge. Down Detector showed that around 11:30 AM Pacific Summer Time, Google Cloud reported over 13,000 incidents, though this number had significantly decreased by early afternoon.

Google stated that due to Identity and Access Management (IAM) issues, multiple GCP (Google Cloud Platform) products were affected, including Gmail, Google Calendar, Google Chat, Google Cloud Search, Google Docs, Google Drive, Google Meet, Google Tasks, and Google Voice. Several Workspace products also encountered service issues.

Google Cloud's status page stated: "Our multiple GCP products encountered service problems" and noted that the service interruption began at 10:51 AM Pacific Summer Time. "Our engineering team continues to investigate this issue."

A few minutes later, Google stated that customers "continue to experience varying degrees of impact", and engineers were working to restore services, but the company had not yet determined an estimated repair time.

After a quick fix, at 12:41 PM Pacific Summer Time, Google stated on the Google Cloud status page: "Our engineers have found the root cause and taken appropriate mitigation measures. While our engineers have confirmed that underlying dependencies have been restored at all locations except us-central1, we note that customers will continue to experience varying degrees of impact across Google Cloud products. All relevant engineering teams are actively involved and committed to service restoration. We do not yet have an estimated time of full service recovery. We will provide updated details before 1:30 PM Pacific Summer Time on Thursday, June 12, 2025."

At 2:00 PM Pacific Summer Time, Google again updated the latest progress on the Google Cloud status page.

Google stated: "We have implemented mitigation measures for this issue in us-central1 and multiple US regions and have seen signs of recovery. We have received confirmation from internal monitoring and customers indicating that Google Cloud products are recovering in multiple regions, with some recovery signs in us-central1 and multiple US regions," Google Cloud said, expecting recovery to be completed within an hour.

At 3:16 PM Pacific Summer Time, Google Cloud stated: "As of 1:45 PM Pacific Summer Time, most Google Cloud products have been fully restored."

However, some products remained unrecovered, including Google Cloud Dataflow: Customers might experience delays when using Dataflow operations due to gradually clearing backlogs; Vertex AI Online Prediction: Customers might continue to encounter increased 5xx errors when using certain models in Model Garden; Personalization Service Health: Delays in Personalization Service Health updates, with recommendations to continue using the Cloud Service Health dashboard for updates.

By 6:27 PM Pacific Summer Time, Google Cloud stated that all services had been restored to normal.

After nearly three hours of major service paralysis, many enterprises were severely impacted by Google Cloud's outage.

Spotify and Cloudflare Become Severe "Disaster Areas"

Specifically, who else was severely affected besides Google's applications and services?

E-commerce software supplier Shopify, a major Google Cloud customer, was one of the most severely affected enterprises. They stated in a post on X, "We are aware of an issue affecting multiple services".

Just after 2:15 PM Pacific Summer Time, Spotify's Down Detector report numbers surged, and problems persisted thereafter, with reports continuing to pour in long after the initial interruption.

From that point, Spotify's access volume steadily declined, and despite Google's platform largely recovering, Spotify's issues took much longer to resolve.

Spotify's paralysis lasted approximately 3 hours, with subsequent decline continuing for several more hours.

Cloudflare was also "severely impacted" by Google.

A Cloudflare spokesperson told media: "This was a Google Cloud outage. A few Cloudflare services use Google Cloud and were therefore affected. We expect them to recover quickly. Cloudflare's core services were not impacted."

Regarding the more specific interruption cause, Cloudflare updated on their status page, attributing the failure of their "critical Workers KV service" to "critical third-party service interruption".

In other words, certain Cloudflare products relying on KV service for storing and propagating information became unavailable due to the service interruption of a critical dependent third party.

Interestingly, Cloudflare released a message about its own issues at 11:19 Pacific Daylight Time, indicating that it had already felt the impact before Google notified its customers.

At 12:12 Pacific Daylight Time, Cloudflare reported that "service is beginning to recover" but warned that "we expect intermittent errors for affected services as systems process retries and caches fill up."

At 1:57 PM Pacific Time (20:57 UTC), Cloudflare stated on its status page that "all Cloudflare services have been restored and are fully operational".

The company said it would continue to "monitor platform metrics to ensure continued stability".

Cloudflare noted that the affected services during the outage included Access, WARP, Realtime, Workers AI, Stream, parts of the Cloudflare dashboard, and AutoRAG.

Cloudflare is a major cybersecurity and content delivery network provider. On Thursday, the company's stock price dropped by 5%.

Just now, Cloudflare CTO Dane Knecht posted an apology on X regarding this interruption and said a comprehensive post-mortem report will be released soon.

"We disappointed Cloudflare customers. Our Workers KV service failed, and downstream products dependent on this service also failed. We will soon release a complete post-mortem analysis report.

I know that such incidents have real and serious impacts on teams working and serving customers worldwide.

In addition to the specific details of this incident, the post-mortem will also explain how we will remove this failure case. We will do our best to correct this issue and sincerely apologize for the problems caused today."

https://www.tomsguide.com/news/live/spotify-down-live-updates-6-12-outage

https://news.ycombinator.com/item?id=44260810

https://www.cnbc.com/2025/06/12/google-cloud-and-other-internet-services-are-reporting-outages.html

https://www.cloudflarestatus.com/incidents/25r9t0vz99rp

This article is from the WeChat official account "AI Frontline", authored by Dong Mei, published with authorization from 36Kr.