Major technology companies, including ride-hailing giant Uber, online productivity startup Notion, and artificial intelligence coding startup Cursor, have confirmed that they suffered data theft following a security breach at Fireworks AI, a platform that provides developers with access to open-weight artificial intelligence models.
The incident has highlighted the escalating risks associated with third-party vendor dependencies in the modern technology ecosystem. According to reports, a spokesperson for Fireworks AI confirmed that the company discovered unauthorized access to several of its core systems, which ultimately impacted customer data and downstream applications connected to its infrastructure.
In an official security update released on November 9 regarding the incident that originally unfolded in October, Fireworks AI detailed the scope of the breach. The company stated that the unauthorized intrusion allowed malicious actors to access resource metadata, billing information—which explicitly did not include payment card numbers—as well as proprietary model weights and datasets. Fireworks AI stated that it immediately began notifying all customers whose data was compromised as part of its remediation strategy.
Explaining the mechanics of the attack, Fireworks AI revealed that a third-party cloud provider hosting its research and development machines experienced a security incident on October 23. By the following day, October 24, the unauthorized third party leveraged that initial compromise to breach Fireworks development machines. The attackers then executed an unauthorized workload that resulted in broader customer data exposure. Fireworks AI detected this anomalous activity on the same day and enacted its formal incident response procedures to implement immediate remediation measures and close known critical vulnerabilities.
The downstream effects of the Fireworks AI breach have varied widely among its high-profile clientele, bringing distinct security challenges to light for each affected organization. Uber, one of the most prominent companies impacted by the incident, is currently in the process of notifying nearly 150,000 drivers that their personal information and driver’s license details were stolen during the breach at the third-party technology provider.
According to reports, the compromised information includes sensitive personal identifiers such as names, physical addresses, dates of birth, and driver’s license numbers. A spokesperson for Uber clarified the situation, noting that the notifications being sent to drivers directly relate to the security incident impacting their third-party vendor, Fireworks AI, which possessed authorized access to specific segments of Uber data for development and operational purposes.
Other technology companies that relied on Fireworks AI have also shared detailed timelines and remediation efforts with their user bases. AI coding startup Cursor published an update on Friday, November 13, explaining that the security compromise began at Fireworks AI when an attacker successfully obtained an access key that Cursor had previously issued to Fireworks. The attacker then weaponized a stolen credential extracted from the exposed data to publish a malicious extension. Cursor noted that it first alerted its customers to the potential exposure on October 30, while continuously monitoring the situation until it observed and successfully blocked the final attacker activity on November 1.
Meanwhile, online productivity platform Notion addressed the incident in a security advisory published on Thursday, November 12. Notion reported that the Fireworks AI security incident impacted less than 1 percent of its total user workspaces. In response, Notion’s security team promptly contacted all customers whose workspaces were determined to have been affected by the downstream breach originating at the AI vendor.
The cascading security failures stemming from the Fireworks AI breach underscore a broader, systemic vulnerability within the technology sector regarding third-party supplier risks. Industry data underscores just how prevalent these supply chain vulnerabilities have become in recent years. Research published by Verizon found that nearly 30 percent of all data breaches recorded over a twelve-year period involved a third-party entity, such as a supplier, vendor, hosting partner, or outsourced IT support provider.
Further compounding these findings, a PYMNTS Intelligence report titled "Vendors and Vulnerabilities: The Cyberattack Squeeze on Mid-Market Firms" highlighted that when cyberattacks target the technology ecosystem, middle-market companies and outside partners frequently find themselves playing the role of unwitting accomplices after their own systems have been covertly compromised by malicious actors seeking broader network access.
As investigations into the Fireworks AI incident continue, affected companies and industry observers alike are re-evaluating the security protocols governing third-party integrations, API keys, and vendor access privileges in an era where artificial intelligence development heavily relies on interconnected, third-party cloud infrastructure.
