The internet and the structures that uphold it prove surprisingly fickle when an app, service or platform you regularly use suddenly goes down. Throughout 2025, Tom’s Guide tracked 75 outages across nearly 50 different services and companies.
Looking back through the various blackouts, some were more impactful than others, with most lasting only an hour or two. To judge how staggering the interruptions were this year, we looked at how many other services were impacted or how often a service went down.
AWS — October 21
One of the largest outages this year took down Amazon Web Services, a cloud platform that acts as infrastructure for apps and websites. Think of it as a crucial backbone of the internet.
The more than 12-hour outage specifically hit the US-East-1 hub, affecting people on the East Coast of the United States and any services based on the infrastructure in that region. In all, more than 1,000 different apps or services were potentially affected. Between first-hand experience and Down Detector reports, we found 48 major services impacted from Amazon itself to Fortnite, Snapchat, Venmo, and Xbox.
In our reporting, we found that a DNS issue inside AWS’ DynamoDB service caused a chain reaction that rippled across the internet. Simply put, one of the internet’s address books forgot where critical servers were located.
ChatGPT — multiple outages
Of all the services we tracked this year, ChatGPT suffered the most interruptions. From minor hour-long drops that we didn’t formally track to longer outages that knocked multiple aspects of OpenAI’s models out. In general, ChatGPT was prone to going down at least once a month for an hour to an hour and a half at a time.
In all, from January to October this year, we tracked 11 outages, two of which were specifically located in the United Kingdom.
The worst one was a global blackout that caused ChatGPT to go offline for over 10 hours on June 10. That outage affected ChatGPT’s free and subscription tiers as well as the image generator Sora. OpenAI confirmed the outage, but we never received an explanation for what caused the severe interruption.
Cloudflare — November 18 and December 5
Like AWS, Cloudflare is a considerable backbone of the internet. When the company experienced a global outage in mid-November, it took down nearly a quarter of all websites.
Cloudflare is a global cloud network that operates multiple websites. The November outage lasted nearly six hours and was reminiscent of the CrowdStrike issue that took down multiple airline and government systems in 2024.
Everything went down from X to Uber Eats, and even the outage tracking site DownDetector crashed.
Cloudflare claims the outage was not caused by a cyber attack or hack. Instead, a blog post said a change to one of the firm’s databases altered permissions that accidentally doubled a file size so much that it caused internal software to fail.
For the second outage, according to Cloudflare’s CTO, the company attempted to fix a security vulnerability in the early morning hours of December 5. Implementing that fix required shutting down some Cloudflare features, which caused the brief outage.
Garmin — January 28
Early in the year, a strange bug affected several Garmin smartwatches that prevented users from accessing GPS, showing only a blue triangle or the Garmin logo. It’s a core feature of many of the best smartwatches, especially for Garmin users who use it to track workouts.
Reports were widespread on January 28, but Garmin managed to issue a fix for most people by the end of the day. However, there were still reports of a blue triangle on January 29.
The bug was so bad that Garmin had to release a support page instructing users on how to reset their Garmin watch. Some devices required a hard reset, while a few could get away with soft resets. We did find that in some cases, just resyncing with the Garmin Connect app was enough. It’s not quite the longest outage we covered this year, but it’s up there.
Google — multiple
A variety of Google services suffered outages this year. Separately, both Google Cloud and its Gemini AI went down for three hours and just about two hours, respectively.
The Cloud outage affected Nest thermostats alongside YouTube and Google Drive, though Nest devices seemed to take the brunt of the interruption, lasting well past the company’s restoration of service.
However, the bigger impacts were felt when Google Drive and Google Workspace went down twice this year. The larger interruption lasted nearly six hours, making services like Sheets, Search, and more unavailable for much of the workday. We never saw an explanation for why, but at one point, Google engineers suggested rebooting routers as a workaround, which seemed to work.
PlayStation Network — February 2 to February 3
Of all the outages we covered in 2025, the PlayStation Network blackout from February 2 was the longest, lasting just over 24 hours in total. The outage started around 6 p.m. ET on Friday and didn’t return to normal until the next Saturday.
Sony blamed an “operational issue” for the second-longest PSN outage ever. It’s not clear exactly what happened, but The Verge’s Tom Warren suspected it was a massive DDoS attack similar to one that hit Xbox over the 2024 holidays.
Sony offered subscribers an “additional 5 days of service.” The outage even affected a Capcom beta test for ‘Monster Hunter Wilds’, though the company extended it by a day for PlayStation players.
Reddit — multiple
After ChatGPT, Reddit and X tied for the second most outages that we covered this year. Reddit went down at least five notable times this year with the most recent kicking off in November.
Most Reddit interruptions only lasted a couple of hours, but the social platform was spotty in April, suffering two outages in quick succession. The larger of the two started on April 21 suddenly crashing around the globe with over 110,000 reports flooding Down Detector in a matter of minutes.
According to a Reddit spokesperson, the outage was caused by a brief inability to connect to one of the platform’s databases. At the same time, we saw an uptick in outage reports for AWS and CrowdStrike, but it’s not clear if those issues were related. The second outage occurred on April 29 but barely reached an hour.
Spotify — multiple
While Spotify’s several outages didn’t last nearly as long as some others, what made the music streamer stand out this year was the amount of reports we saw. Two different outages spiked nearly 50,000 reports. Nothing comes close to the amount that Reddit received, and even the 24-hour PSN outage never saw numbers that high.
To be fair, one of the four outages we covered was related to the Cloudflare blackout. Obviously, people care very deeply about having access to their music.
The next time Spotify takes a set break, why not check out the best music streaming services that aren’t the green and black.
X (nee Twitter) — multiple
As mentioned, X tied with Reddit for the second most outages we covered in 2025, with five total. The most recent knocked the former bird site down in November, but it was related to another Cloudflare issue.
The biggest outage occurred in May, when a data center in Oregon caught fire and knocked portions of the platform offline for multiple days. Unlike the 24-hour PSN outage, which affected every PlayStation user, service for X was intermittently interrupted, so not everyone felt the effects of the blackout. However, for some, the outage lasted from Thursday, May 22, all the way through Saturday, May 24.
Twitter is usually extremely opaque when it comes to reporting on its service or what may be causing issues. So it was a surprise that the company even acknowledged the data center fire. As we covered the outage, the X engineering account made one tweet acknowledging people were facing issues, and only one other tweet about the incident afterwards.
Payment services — multiple
It’s not exactly a singular service, but throughout 2025, we covered outages that hit multiple different payment services including Apple Pay, PayPal, Venmo and Zelle.
I’ve included the payment services here because of their impact. Many people have transitioned to sending and receiving money and making payments via services like Zelle or Apple Pay. Digital banking is on the rise but if you can’t access your funds, it becomes extremely frustrating.
As one of the more family-friendly comments read during one of the PayPal outages, “PayPal needs to figure this out, it's BS quite literally manipulating people's money and their lives.”
“PayPal needs to figure this out, it's BS quite literally manipulating people's money and their lives.”
Affected user
Most of the outages lasted only a couple of hours, but Zelle took a huge hit in May when a third-party payment infrastructure company, FiServ, suffered an “internal error” that broke online transactions for several banks, including Bank of America and Ally. Functionally that meant Zelle transfers couldn’t be made in or out of accounts connected to affected banks.
That outage lasted multiple hours and wasn’t truly resolved until the next day when it appears backlogged payments finally went through.
As the world grows ever more connected, especially via AI services, expect outages to continue. Tom’s Guide will be there to keep track and help keep you abreast of what services are down and how you can find alternatives and workarounds to go about your digital life.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
More from Tom's Guide
- Why I don't use Spotify — and the alternative I use instead
- Tired of Google Maps and Apple Maps? Here are your best alternatives
- I ditched the iPad mini for an Android alternative — and this one feature means I might not go back
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
