Two significant technology outages disrupted major online services this week, with OpenAI’s ChatGPT suffering an eight-hour downtime on June 10, followed by Google Cloud infrastructure problems affecting multiple platforms on Thursday.

OpenAI’s popular AI assistant ChatGPT and image generator Sora experienced their longest outage to date, lasting more than eight hours, beginning shortly before midnight PDT on June 10.

Reports on Downdetector initially spiked around 12am PDT, with the most significant problems occurring around 3am and 6am PDT.

OpenAI acknowledged the issue at 7:16am PDT, confirming recovery efforts were underway by 8am PDT, though the company warned full restoration could take hours.

By 8:42am PDT, Sora had been restored, and ChatGPT was approaching full recovery.

The extended downtime surpassed previous ChatGPT outages, including a four-hour incident in December 2024 and a five-hour disruption in June 2024.

The pattern of major outages occurring approximately every six months is likely to prompt an internal review at OpenAI.

On Thursday afternoon, Google Cloud experienced widespread infrastructure problems beginning around 1:51pm ET, disrupting services including Google Home/Nest devices, Spotify, and numerous other platforms relying on Google’s cloud infrastructure.

Users attempting to access Spotify encountered error messages, including “Audiences in Jwt are not allowed” on the website, while the mobile app displayed connection failures.

The streaming service, a known Google Cloud customer, has previously been affected by similar infrastructure outages.

The Google Cloud issues triggered secondary problems across multiple services, with outage reports spiking for Twitch, Snapchat, Anthropic, Shopify, Discord, and Replit.

Cloudflare also reported outages, though spokesperson Alexander Modiano clarified these resulted from Google Cloud dependencies rather than core Cloudflare service failures.

Google’s status page initially showed no issues but was later updated to report “Multiple GCP products are experiencing impact due to Identity and Access Management Service Issue.”

Public acknowledgment came after Replit CEO Amjad Masad tweeted about the outage at 2:34pm ET.

Google reported at 6:16pm ET that “Most of the Google Cloud products are fully recovered,” though some residual impact remained.

Cloudflare simultaneously listed its related outage as fully resolved.

The consecutive outages highlight the interconnected nature of modern internet infrastructure, where single points of failure can cascade across multiple services and platforms.

Both incidents underscore the critical importance of cloud infrastructure reliability as businesses and consumers increasingly depend on online services for daily operations.

The timing of both outages within days of each other emphasises the vulnerability of centralised cloud services and the potential for widespread disruption when major providers experience technical difficulties.