Story Railway

Railway Suffers Major Security Breach as Hackers Exploit AI Credentials

Cybersecurity
The brief

Railway, a cloud hosting provider, recently confirmed a significant security breach that resulted in approximately $740,000 in fraudulent AI usage charges. The incident, disclosed on October 22, stemmed from a nearly 26-hour battle against attackers who targeted AI provider credentials, which are increasingly seen as valuable assets for cybercriminals. The attackers reportedly gained access to around 100,000 workspace credentials, utilizing them to generate charges for AI services that customers had not requested or authorized. This breach highlights the growing trend of so-called LLMjacking, where AI credentials are becoming the new currency for cybercrime, as noted by David Shipley of Beauceron Security.

The attack began on October 20 when hackers exploited an unknown vulnerability within Railway's build infrastructure, allowing them to bypass filesystem isolation protections. They obtained internal credentials that facilitated access to stored build histories, from which they retrieved a copy of Railway's internal source code. Following this, the attackers exploited another vulnerability related to an improperly configured trust relationship, further compromising Railway's infrastructure and enabling them to decrypt customer environment variables. Throughout the incident, Railway's engineering team worked around the clock to mitigate the threat, which involved a complex series of exploits and lateral movements across multiple systems.S1S2

Railway's response included hardening their defenses in real-time as the threat unfolded, emphasizing that this was not a straightforward breach but rather a dynamic contest between attackers and defenders. By October 21, the company reported that the attackers had been fully evicted from their systems. However, the damage was significant, with the breach exposing linked-account tokens belonging to more than 1,850 users, including GitHub and OAuth tokens, which Railway has since reset. In the aftermath, the company is working to reimburse affected customers, including for charges that service providers may refuse to refund.S1S2

Timeline · newest first
2h Computerworld

Hackers stole AI credentials to get $740,000 in free compute; Railway picked up the bill

As AI model training and distillation increase, AI tokens are becoming a piping hot currency for threat actors. Cloud hosting provider Railway found out the hard way; it has confir…

3h Csoonline

Hackers stole AI credentials to get $740,000 in free compute; Railway picked up the bill

As AI model training and distillation increase, AI tokens are becoming a piping hot currency for threat actors. Cloud hosting provider Railway found out the hard way; it has confir…

Railway Suffers Major Security Breach as Hackers Exploit AI Credentials

2 outlets 2 reports
The brief

Railway, a cloud hosting provider, recently confirmed a significant security breach that resulted in approximately $740,000 in fraudulent AI usage charges. The incident, disclosed on October 22, stemmed from a nearly 26-hour battle against attackers who targeted AI provider credentials, which are increasingly seen as valuable assets for cybercriminals. The attackers reportedly gained access to around 100,000 workspace credentials, utilizing them to generate charges for AI services that customers had not requested or authorized. This breach highlights the growing trend of so-called LLMjacking, where AI credentials are becoming the new currency for cybercrime, as noted by David Shipley of Beauceron Security.

The attack began on October 20 when hackers exploited an unknown vulnerability within Railway's build infrastructure, allowing them to bypass filesystem isolation protections. They obtained internal credentials that facilitated access to stored build histories, from which they retrieved a copy of Railway's internal source code. Following this, the attackers exploited another vulnerability related to an improperly configured trust relationship, further compromising Railway's infrastructure and enabling them to decrypt customer environment variables. Throughout the incident, Railway's engineering team worked around the clock to mitigate the threat, which involved a complex series of exploits and lateral movements across multiple systems.S1S2

Railway's response included hardening their defenses in real-time as the threat unfolded, emphasizing that this was not a straightforward breach but rather a dynamic contest between attackers and defenders. By October 21, the company reported that the attackers had been fully evicted from their systems. However, the damage was significant, with the breach exposing linked-account tokens belonging to more than 1,850 users, including GitHub and OAuth tokens, which Railway has since reset. In the aftermath, the company is working to reimburse affected customers, including for charges that service providers may refuse to refund.S1S2

Timeline · newest first

Sources · 2 citations

S1 Computerworld 1 report · EN
S2 CsoonlineLead 1 report · EN