Northflank: Security Breach Exposes Customers' AI API Keys to Unauthorized Use

Cloud application hosting platform Northflank disclosed a security incident on October 30, 2026, warning all customers to rotate their environment variables and API keys after the cloud security company Wiz alerted it to a threat, and confirmed that multiple customers have reported unauthorized use of AI API keys stored on the platform. According to BleepingComputer, Northflank told customers to assume their secrets had been stolen: "Assume all environment variables have been decrypted. Disable/rotate all keys." The company also stated: "Users have reported unauthorised use of AI API keys supplied to the Northflank platform."

What was the specific vulnerability?

The specific vulnerability is a breach of Northflank's platform that exposed customers' encrypted environment variables, the containers holding database passwords, cloud credentials, authentication tokens and API keys, to decryption and theft. Northflank's guidance to customers was unambiguous: "Assume all environment variables have been decrypted. Disable/rotate all keys," per the October 30 update cited by BleepingComputer.

The company has not disclosed how attackers gained access, how many customers were affected, or whether credentials other than AI API keys were stolen or abused. BleepingComputer contacted Northflank but received no response. Northflank's own status page notes that the customer reports "may be consistent with the broader pattern of attacks targeting AI credentials" and that similar IP addresses appear in Wiz's "Black October" report, which documents a wave of incidents against AI developer platforms.

How was the incident detected?

The incident was detected externally by Wiz, which sent Northflank a threat alert at 11:43 UTC on October 30, 2026. Northflank's published incident timeline, hosted on its status page, shows engineers reviewed the alert one minute later at 11:44 UTC, and that an initial network policy patch went out at 12:02 UTC, roughly 19 minutes after the alert arrived. Additional patches followed at 12:50 UTC and 13:05 UTC, and Northflank requested support from Wiz's Computer Incident Response Team (CIRT) at 13:00 UTC. The CIRT is Wiz's dedicated incident response unit, and its involvement indicates Northflank treated the alert as an active intrusion rather than a routine threat notification. The full timeline is documented on the Northflank status page.

What is the current impact?

The current impact is that every Northflank customer has been told to treat their stored secrets as compromised, and multiple customers have confirmed that attackers used their AI API keys without authorization, creating the risk of unauthorized AI charges on their accounts. As remediation, Northflank enabled credential rotation for PostgreSQL, MongoDB, MySQL and Redis add-ons on October 30, and extended rotation to Minio and RabbitMQ add-ons in October 31 updates, according to the company's status page. It also enabled audit logs for all customers, including free accounts, a step that gives customers visibility into access to their services during and after the incident window.

The incident remains listed as "Identified / Major outage" on Northflank's status page, with seven affected components and an outage window listed from 9:40 AM to 7:00 PM on October 30, 2026, per the status page timeline. The broader context is what Wiz calls "Black October for AI Developer Platforms": at least nine security incidents in the month affecting AI development platforms including Northflank, Railway, OpenCode, Warp and Lightning AI. Per BleepingComputer's summary of Wiz's findings, Railway involved potential exposure of decrypted customer environment variables, OpenCode disclosed exposed AI API keys, and Warp reported a breach exposing AI agent logs and user telemetry data. Wiz says attackers in some incidents specifically targeted environment variables holding AI service credentials for "LLMJacking," the practice of using stolen credentials to access AI models on another organization's keys, generating large API charges or reselling access. In two Wiz-investigated cases, compromised enterprise credentials drove approximately 2.5 million Amazon Bedrock API calls, generating approximately $100,000 in AI usage.

What are the next steps?

The next steps are customer-side: Northflank has instructed all customers to disable and rotate every key and secret stored on the platform, and the company has enabled rotation tooling and audit logging to support that process. Customers of the PostgreSQL, MongoDB, MySQL, Redis, Minio and RabbitMQ add-ons can rotate their credentials through the mechanisms Northflank switched on between October 30 and October 31. Northflank has not yet published an explanation of the attack vector, a count of affected customers, or a determination of whether non-AI credentials were abused, and the incident remains open on its status page. Customers with AI API keys on the platform face the most immediate financial exposure, given the confirmed unauthorized use reported by multiple users and the LLMJacking pattern Wiz has documented across the month's incidents. Until Northflank publishes a root-cause analysis, the platform's own status page remains the primary channel for updates on the investigation.

By the numbers

Source snapshot