# Data Breach at Uber Affects Nearly 150,000 Drivers Due to Third-Party Compromise

*Published November 18, 2026*
*Source: [https://therecord.media/uber-says-150000-impacted-third-party-breach](https://therecord.media/uber-says-150000-impacted-third-party-breach)*

## Executive Summary

Uber has reported a data breach affecting 149,346 drivers due to compromised credentials at a third-party service provider, leading to unauthorized access to sensitive personal information. This breach is part of a broader pattern of attacks linked to AI infrastructure vulnerabilities.

## Article

Uber has alerted regulators across several U.S. states about a recent data breach that exposed the personal information of 149,346 drivers. The rideshare giant was informed by a third-party service provider on October 9 that unauthorized access had occurred using stolen internal credentials. This breach allowed hackers to access images of driver’s licenses along with sensitive personal data such as names, addresses, dates of birth, and license numbers. The compromised data was stored in a cloud environment specifically set up by Uber to facilitate the provider's operations. 

Upon learning of the breach, Uber swiftly terminated its data connections with the affected service provider and has been conducting a thorough investigation. The company has also taken steps to delete or rotate relevant credentials and access details. Despite the breach, Uber's cloud security management protocols were not breached since the unauthorized access involved valid credentials, thus not triggering any security alerts. Importantly, the breach was confined to the provider's system and did not extend to other Uber systems.

The breach is connected to a broader attack on Fireworks, a company specializing in AI infrastructure. Fireworks disclosed that cybercriminals had stolen billing and AI-specific data in a series of breaches last month. This incident is part of a wider pattern of cyberattacks identified by Google Threat Intelligence Group and Mandiant, which have noted multiple occurrences of a tactic called 'LLMJacking' where hackers exploit access to large language models. Several other companies, including Notion and Cursor, have reported being affected by the Fireworks incident as well.
