Uber is notifying 149,346 people, mostly drivers, that their personal and driver’s licence information was stolen. The breach happened at Fireworks AI, a cloud platform for running open-source AI models. The same incident has now affected at least three of Fireworks’ customers.
Valid credentials, no alerts
According to SecurityWeek, Uber learned of the incident on October 9 and immediately cut its data connections with the provider. By October 13, its investigation found that attackers had used the provider’s compromised credentials on October 7 to access data Uber had shared with it.
The stolen data included images of driver’s licences, names, addresses, dates of birth, licence numbers and government ID numbers. Uber said the data sat in an environment its own Cloud Security team created and managed. Alerts were switched on, but none fired because the attacker used valid credentials.
“The service provider’s compromised credentials did not provide access to other Uber systems. The investigation remains ongoing,” the company said. Uber has filed notices in several US states and is offering free credit monitoring and fraud help.
Uber did not name the vendor in its filings. A spokesperson confirmed to TechCrunch that it was Fireworks AI, TechCrunch’s report said. Fireworks confirmed unauthorised access to some of its systems and blamed a financially motivated group that sells stolen access tokens for large language models.
Cursor and Notion also hit
Fireworks first said it had found no evidence that customer personal data, models or datasets had been accessed. In a November 9 update, it said account and billing information, some customer datasets and incomplete models were affected, SecurityWeek reported.
Kaseya’s weekly breach roundup noted that AI coding company Cursor confirmed attackers used stolen Fireworks credentials to push a malicious Claude Code extension through its marketplace. Notion said about 1% of its workspaces were affected.
Uber has said it has found no misuse of the data so far, eBuilder Security reported. However, licence images combined with names, addresses and dates of birth are valuable to fraudsters, who can use them to open accounts, pass identity checks or make phishing messages more convincing.
Why this breach matters
The attackers never broke into Uber itself. They logged in as a trusted supplier, which is why Uber’s own monitoring stayed silent. As more companies send customer data to outside AI providers for model training, document processing or support automation, each provider becomes another door into that data.
Security teams can reduce the risk by:
- Sharing only the minimum data with each vendor and deleting it when the job is done.
- Limiting vendor credentials to specific networks and short lifetimes.
- Alerting on unusual data volumes or access times, not just failed log-ins.
- Writing breach-notification timelines into vendor contracts.
What it means for African readers
African banks, fintechs and ride-hailing firms are adopting AI tools fast, often through foreign cloud vendors. Licence and ID images are exactly the kind of know-your-customer data these firms hold.
Nigeria’s Data Protection Act, Kenya’s Data Protection Act and South Africa’s POPIA all hold the company that collected the data responsible, even when a supplier is breached. Firms should check where their customers’ ID documents travel, and customers should watch for phishing that uses leaked personal details.








