Service disruptions inhibited many Google users' AI work on Wednesday. Here's what Google said happened. Katelyn Chedraoui See full bio Katelyn Chedraoui June 11, 2026 6:26 a.m. PT The Gemini outage is mostly resolved, but could it happen again? Google's Gemini outage, which prevented people who use Google apps and services from working, should be over now. The company confirmed at 6 p.m. PT/9 p.m. ET on Wednesday that all affected users should be good to go in a dashboard status page. Gemini users got an unpleasant surprise on Wednesday when the company's AI assistant wasn't working properly. DownDetector received over 1,600 reports of malfunctions during peak work hours. (DownDetector is owned by the same parent company as CNET, Ziff Davis.) Those reports fell throughout the day as Google rolled out fixes. It's unclear how many Google users were affected. The issue was limited to Gemini in Google Workspace, which includes Drive, Docs, Sheets, Slides, and the Gemini app. The service disruption is affecting people on MacOS, web, iOS, Gemini in Chrome and Android. Errors are popping up with codes 1099 and 1076, saying, "Something went wrong." After touting Gemini at last month's Google I/O, this week's service disruption made it impossible for many people to use Gemini. Throughout yesterday's Gemini outage -- one of the biggest AI-related outages in Gemini's history -- we didn't know what was causing it. Google shed a little bit of light on the subject in its final status update. "From preliminary analysis, the issue was triggered by a performance issue in our backend database which impacted the retrieval of Gemini App tools catalog," the company wrote. In layman's terms, that means a behind-the-scenes issue made it impossible for Gemini to pull the right information to complete tasks. Google engineers fixed the issue by optimizing the load distribution across the backend database. CNET reached out to Google for additional information on what caused this and what, if anything, will be done to prevent a future outage. Google declined to comment. Google still notes on its Workspace Status Dashboard that the service disruption is ongoing, with a few reports still trickling in on DownDetector, though down drastically from earlier in the day. Google says its next update will arrive at 8:30 p.m. PT, but maintains that most of its users are no longer affected by the disruption. MacOS is one of the affected platforms of the Gemini outage. And just as folks like me on the east coast are wrapping up their work day, Google says its mitigation measures have worked well enough that the "majority of users" should be, in the words of Matthew McConaughey, alright, alright, alright. "Our engineering team has applied mitigations to reduce impact and confirmed the majority of users are no longer observing impact," Google wrote. "We will continue to monitor the service for stability." Still no word on what caused the issue. West coast residents, you should be back in action for the rest of the day. (Sorry, if you were enjoying an AI-free work day.) Claude has some minor service issues on Wednesday, too. On a rough day for AI users, Anthropic reported some issues with its AI chatbot, Claude, too. On a status update page, the company said it had resolved issues with its Haiku 4.5 model after reports of malfunctions. Those issues should have been resolved by 10:21 a.m. PT, according to Anthropic. It isn't clear what caused the problems. Anthropic dropped the first of its much-awaited Mythos-level AI models on Tuesday. It's called Fable 5, and if you pay for Claude, you can use it now. Anthropic did say that it anticipates that demand will be very high, so it's possible Claude users may encounter more access issues in the near future. Guess it's a good thing for AI enthusiasts that OpenAI's ChatGPT has had a quiet day? Fixes are on the way for affected Google users. Google confirmed in an update at 11:37 a.m. PT that it's rolling out fixes and that there are signs of recovery. "Our engineering team has applied mitigations to reduce impact and continues to investigate the root cause," Google wrote. "We are seeing signs of recovery and will continue to monitor progress." Reports submitted to DownDetector have steadily dropped since its peak of over 1,600 incidents of Gemini malfunctioning earlier in the day. But because Google has millions of users, it's very likely that many people are still encountering issues using Gemini in their Workspace apps. Heads up: Gemini is currently experiencing an outage. We're on it and will get everything back up ASAP. Some of the fixes are in, the rest coming very soon. Stay tuned for updates, and thanks for bearing with us! Gemini App Vice President Josh Woodward confirmed in an X post at 10:31 a.m. PT that some fixes to the Gemini outage are on the way. Google just had its I/O developers conference, where Gemini AI was everywhere in Google's business products. A "mitigation" or fix is on the way, Google confirmed at 9:24 a.m. PT. There's no ETA on when it will be deployed. Google still hasn't given us many details about the underlying issue, but it did confirm that the error codes 1099 and 1076 are part of the issue. I saw these error codes pop up in the bottom-left corner of my screen, though they may vary depending on the app you're using. Keep scrolling to see what apps and platforms are affected. There is still no workaround or temporary fix. CNET uses Google Workspace, so I went hunting through my Drive, Docs and Sheets earlier this morning to see if I was one of the affected customers. Now my time as "marked safe" is over, so I can share what the error looks like. I first encountered the error when I asked Gemini (through Chrome on my Mac) to analyze the contents of a PDF that I uploaded. I had the flash model selected, which I immediately knew wouldn't be the best model for the task after I hit "run." But I didn't have to wait long to fix my error, because Gemini couldn't run the task at all. It ran for a second or two before reverting me to the prompt screen. In the lower-left corner, error code 1099 popped up, too quickly for me to screenshot. I had to see if it was a fluke. I started a new chat with a seemingly easy web search for Gemini. The same thing happened -- it tried to run for a second before almost immediately quitting. This time, it sent me back to the main homepage and had a different error code, shown above. As of now, there's no way to work around this issue. Luckily for me, these are both tasks that I don't necessarily need Gemini to assist with. Google Drive, Docs, Sheets and Slides are all part of Google Workspace. In an update at 8:10 a.m. PT, Google said that the Gemini app in Workspace -- which includes Drive, Docs and Sheets -- is malfunctioning on the following platforms: These are the platforms that are displaying the "something went wrong" error. There's still no workaround. Google said it will issue another update at 9:30 a.m. PT/12:30 p.m. ET. Google Health uses a chatbot-like interface to give you advice based on your biometric data. Google has put its Gemini AI into seemingly as many of its products as it can. So when I see something like "Gemini is down," my mind goes to all of the Google products that have integrated the AI system over the past few years. Today's issue isn't widespread; according to Google, it's affecting only Workspace users. I checked my Google Health app, which is what the old Fitbit app has become, and had a conversation with the chatbot there about why the fact that I feel like I'm coming down with a cold isn't showing up in my biometric data. Was this helpful information? As Gemini told me, I shouldn't trust it for medical advice, so not really. But it did generate responses. As companies like Google integrate AI into more and more of their products, it does create another potential source of breakage. Google hasn't yet said what the issue with Gemini is, but for Workspace users, it's resulting in errors that read, "Something went wrong." Google is categorizing it as a service disruption, not an outage. There are no workarounds as of 8 a.m. PT. Most likely, the Google engineering teams caught the issue late last night or early this morning, before many of us logged on for work at 8 a.m. They're currently heads down trying to understand what the problem is and how to fix it. We'll be here tracking how the issue is affecting you and when it is eventually resolved.