Railway halts every deployment after intruders may have accessed customer secrets — AI inference keys appear to be the target

• Who: Railway, the cloud deployment platform led by CEO Jake Cooper.
• What: an unauthorized third party may have accessed encrypted customer environment variables, and obtained the decrypted contents of some of them.
• When: detected Wednesday morning, October 21, 2026; disclosure published at 4:44 p.m. EDT (20:44 UTC).
• Secrets at risk: API keys, database and registry credentials stored as environment variables.
• Response: all deployments paused; every user urged to roll all env vars and private registry credentials and review service logs; confirmed affected accounts to be contacted by email.
• AI angle: Railway says the intrusion appears part of a series of attacks targeting platform companies to compromise AI inference providers; upstream providers were notified to revoke keys proactively.

Cloud deployment platform Railway is telling every customer to rotate their secrets after detecting an intrusion on Wednesday morning in the systems that store customer environment variables — and it believes the attack is part of a campaign aimed at AI inference providers.

What happened

Railway says it detected anomalous activity early Wednesday in the systems that store customer environment variables. After further investigation, the company says an unauthorized third party may have accessed encrypted environment variables — and, for some of them, obtained the decrypted contents. Those variables can hold API keys as well as database and registry credentials, making them the keys to customers' production infrastructure. The disclosure was published at 4:44 p.m. EDT, and updated later in the day as the investigation continued. (Primary source: Railway's security disclosure)

The target: AI inference providers' keys

Railway's reading of the incident reaches beyond its own platform. The company says it believes the intrusion is part of a series of attacks recently targeting platform companies “with the aim of compromising AI inference providers.” It has notified those upstream providers to revoke all keys proactively. In other words, the intruders may not have been after Railway's customers at all — they were after the AI API keys that live inside customers' deployments. This is not Railway's first security disclosure: the company also published an incident report in December 2025 covering a large-scale supply-chain attack on its platform.

Everything frozen — including the fix

All deployments are paused while Railway investigates. Until builds are restored, customers cannot roll out new versions — which creates an awkward bind: Railway is asking users to rotate compromised credentials, yet the deployment freeze means the updated applications that would carry the new credentials cannot ship yet. Railway says it will post updates to its disclosure and to status.railway.com, and will email affected customers directly once it confirms which accounts were hit. (RuntimeWire)

What Railway wants you to do

Railway's initial analysis suggests the impact may be limited to a subset of users — but it is recommending that all users immediately roll every environment variable stored in Railway, including any API keys, as well as any private registry credentials, and review logs on those services for unusual activity.

Why it matters

The incident lands at an awkward moment for the AI boom: platforms like Railway are where thousands of developers host AI applications, and the API keys that unlock frontier models live exactly where Railway says the intruders went looking. It also exposes a structural risk in the platform pitch — a cloud that promises to remove infrastructure work can, in a crisis, ask customers to do the most urgent infrastructure work themselves while blocking the deployments that would finish the job.