A cyberattack on cloud hosting provider Railway has highlighted the growing risks linked to stolen AI credentials. Attackers allegedly accessed customer secrets and used compromised AI provider keys to generate around $740,000 in unauthorised usage charges.
Railway said the incident began on October 20 and involved a nearly 26-hour battle with attackers. The hackers reportedly exploited vulnerabilities in its build infrastructure, accessed internal systems and obtained credentials used to decrypt customer environment variables.
The company said attackers accessed a subset of approximately 100,000 customer variables, representing up to 1.85% of its workspaces. They also exposed linked-account tokens belonging to more than 1,850 users, including GitHub and OAuth credentials. Railway worked with GitHub, GitLab, Anthropic, OpenAI, OpenRouter, SpaceXSI, Stripe and other providers to revoke potentially compromised keys.
How stolen AI keys were exploited
Railway said attackers searched its databases for AI provider credentials before targeting other cloud, database and payment services. The stolen keys were reportedly sold to AI access providers for activities such as model distillation, which transfers knowledge from larger models to smaller ones.
Customers found that their credentials had been used for coding queries and training-data tasks they had not initiated. Railway said it is working to reimburse affected customers, including charges that AI providers refuse to refund.
The company disclosed the incident on October 22 and advised users to rotate sensitive environment variables stored before October 21. It also reset exposed linked-account tokens, replaced internal keys, moved to new infrastructure and introduced tools for mass revocation of app permissions.
Security experts warned that stolen AI credentials are becoming valuable targets. “AI credentials are becoming the new credit card numbers,” said David Shipley of Beauceron Security.
Frank Dickson of Dickson Research compared the practice to “copying a textbook one page at a time, using library cards that belong to other people.” He said attackers distribute millions of AI requests across legitimate customer accounts, leaving the account owners with the bills.
Customers have been advised to rotate AI, cloud, database and payment credentials, review account and usage logs, restrict database access, enable spending limits and report unauthorised charges. Experts also recommend short-lived credentials and logging every decryption.
Also read: Viksit Workforce for a Viksit Bharat
Do Follow: The Mainstream LinkedIn | The Mainstream Facebook | The Mainstream Youtube | The Mainstream Twitter
About us:
The Mainstream is a premier platform delivering the latest updates and informed perspectives across the technology business and cyber landscape. Built on research-driven, thought leadership and original intellectual property, The Mainstream also curates summits & conferences that convene decision makers to explore how technology reshapes industries and leadership. With a growing presence in India and globally across the Middle East, Africa, ASEAN, the USA, the UK and Australia, The Mainstream carries a vision to bring the latest happenings and insights to 8.2 billion people and to place technology at the centre of conversation for leaders navigating the future.