Uber has confirmed that a security breach at an unnamed third-party service provider resulted in the exposure of personal information for 149,346 individuals across the United States. The rideshare company began notifying state regulators after learning on October 9 that internal credentials belonging to the vendor had been stolen and subsequently used to gain unauthorized access to an Uber-managed cloud storage environment.
According to the company, the compromised data included images of driver’s licenses, as well as extracted text containing names, physical addresses, dates of birth, and driver’s license numbers. Uber stated that the affected cloud storage environment had been specifically established to facilitate the service provider's work. While the company maintained that this environment was subject to its own internal cloud security management processes, the breach remained undetected because the attackers utilized valid credentials to gain entry.
Uber has reported that it immediately terminated all data connections to the service provider upon discovering the intrusion. Furthermore, the company has taken steps to rotate or delete all credentials and access information previously associated with that vendor. Uber emphasized that the scope of the unauthorized access was limited to the specific environment used by the provider and did not extend to the broader Uber corporate network or other internal systems.
While Uber has not publicly identified the service provider involved, reports indicate the incident is linked to a recent cyberattack on the AI infrastructure firm Fireworks. Fireworks issued an advisory acknowledging that attackers stole billing information and AI-specific data during two separate incidents in October. Following the disclosure by Fireworks, other organizations, including Notion and Cursor, have also confirmed they were impacted by the same series of events.
The breach is reportedly part of a broader trend of attacks involving the hijacking of enterprise access to large language models, a tactic security researchers refer to as LLMJacking. Analysts at Google Threat Intelligence Group and Mandiant have observed a surge in these incidents throughout 2026. These campaigns have targeted a wide range of organizations across the technology, financial services, retail, and transportation sectors, affecting entities in North America, South America, and the Asia-Pacific region.
The security community has noted that LLMJacking exploits the way organizations integrate third-party AI services into their workflows. By compromising the credentials used to interface with these models, attackers can bypass traditional security perimeters. In this instance, the use of legitimate, albeit stolen, credentials allowed the unauthorized parties to extract sensitive data without triggering the standard alerts that might have otherwise flagged the activity.
Uber’s investigation into the incident remains ongoing. The company has not provided further details regarding the specific nature of the service provider's relationship with its AI infrastructure or why the credentials were able to access the specific cloud storage environment containing driver information. As the investigation continues, the company is working to notify affected individuals and coordinate with relevant regulatory bodies.





