Railway, a cloud platform that spends millions each month on Google Cloud, suffered a major 8-hour outage after Google’s automated systems mistakenly suspended its production account. No warning. The founder says he’s “gobsmacked,” and is demoting GCP to backup-only.
Railway is a large cloud platform with over 3 million users that powers live apps and businesses. But on Tuesday night, suddenly everything went dark. Customers flooded social media with frustration over lost clients and revenue.
The company quickly acknowledged the issues.
“The Railway dashboard is currently unavailable, and all running Railway services are down. We're working with our upstream provider to restore service,” Railway posted on May 20th, 2026.
For 8 hours, between 22:20 UTC May 19th and 06:14 UTC on May 20th, the cloud platform powering around 10 million services experienced a platform-wide outage.
The root cause? Google Cloud blocked its account.
“I am gobsmacked that a company could literally shut down with no notice, an account that spends $2 million a month with tens of millions of lifetime spend. I'm bewildered, frankly,” posted Jake Cooper, founder of Railway.
The incident jolted developer communities. It is not the first time Google’s mistake has disrupted a major customer, renewing concerns over trusting automated systems for high-stakes accounts.
What happened?
Railway uses Google Cloud Platform (GCP) for its dashboard, API, and parts of its network infrastructure, the company shared in a released postmortem.
When GCP suspended Railway's production account, it instantly pulled the company’s GCP-hosted API, control plane, databases, and compute infrastructure offline.
“Users immediately experienced 503 errors on the dashboard and API, including ‘no healthy upstream’ and ‘unconditional drop overload’ messages, and were unable to log in. All workloads hosted on Google Cloud compute were taken offline,” the incident report reads.
It took only 9 minutes after the first API health check failures to identify the root cause, since Google Cloud Platform suspended Railway’s production account. A ticket was immediately filed, and 10 minutes later, the account access was restored. However, servers remained offline, and storage was still inaccessible.
While other workloads continued running for a while on the company's own Railway Metal and AWS platforms, the incident quickly cascaded out of control. Railway uses GCP to direct traffic, and as cached network routes expired, other services became unreachable, too.
“Railway Metal and AWS began returning 404 errors as the networking could no longer resolve routes,” the report reads.
“At peak impact, all Railway workloads across all regions were rendered unreachable.”
It took hours to recover layer by layer: storage, networking, servers, deploys, and services.
A secondary hit came when GitHub began rate-limiting Railway’s authentication (OAuth) and webhook integrations – some users once again were unable to log in, and builds were blocked again.
“Google Cloud placed Railway’s production account into a suspended status incorrectly, as part of an automated action,” the company said.
The report also mentions that GCP’s automated action extended to many other accounts, and was platform-wide – there was no proactive outreach to customers prior to the restriction.
“Deepest apologies,” the CEO of the company kept posting while the incident was ongoing.
“It’s so beyond insane. I'm trying to stay composed.”
Cybernews has reached out to Google for a comment and will update the story with its response.
Strategic shift away from GCP
“The company and its CEO accept the architectural blame over single provider action cascading into a full platform outage, and it was their own design decision.
Moving forward, the company said it plans to remove Google Cloud services from the “data plane’s hot path” entirely and keep them only for secondary/failover. Other architectural changes will ensure that core services, especially user-facing components, are not dependent on any one vendor or platform.
“All Railway users are entitled to, and should expect, reliability. We chose GCP originally. We moved the vast majority of workloads off. We will now only run it as redundant replicas (not primaries),” Cooper said.
The incident has sown distrust and sparked industry-wide discussions on social media. It’s not the first incident affecting large Google customers – just two years ago, Google deleted the account of Unisuper, an Australian retirement savings fund.
Some developers warn that Google routinely suspends accounts through opaque automated systems with no human review, no warning, and no explanation, and Railway is just a “big enough name” to top the Hacker News main page.
“This should be a warning to anyone running GCP. They suspend accounts left, right, and center without even thinking about what they're doing. It seems like they use Gemini 3.1 Pro to run their production decisions,” one of the comments on Hacker News reads.
Until Google issues an official statement, the reason for the account suspension remains unknown, fueling speculations.
Unlock more exclusive Cybernews content on YouTube.
Your email address will not be published. Required fields are markedmarked