Northflank Breach Leads to AI API Key Abuse

Northflank, a cloud platform for deploying apps, databases and AI workloads, told customers on Thursday, Oct. 30, to assume every environment variable stored on its service had been decrypted and to rotate all keys. Late that night, the company said users had reported unauthorized use of AI API keys they had supplied to the platform, according to its status page.

Northflank tied the activity to a broader pattern described by cloud security firm Wiz, which counted at least nine security incidents at AI developer platforms in October alone.

Key takeaways

• Northflank received a Wiz threat alert at 11:43 UTC on Oct. 30 and told users about two hours later: "Assume all environment variables have been decrypted."
• By 23:38 UTC, users had reported unauthorized use of AI API keys stored on the platform. Northflank has not disclosed the entry point or how many customers were affected.
• Wiz says Northflank is the ninth AI developer platform hit in October. In two cases it studied, stolen keys drove about 2.5 million Amazon Bedrock calls.

Northflank security incident timeline

Northflank's status page lays out a fast response. Engineers reviewed the Wiz alert at 11:44 UTC, deployed an initial network policy patch at 12:02 and pushed more patches at 12:50 and 13:05. At 13:00 the company asked Wiz's incident response team for help.

At 13:58 UTC, Northflank moved the incident to "identified" and told customers: "Assume all environment variables have been decrypted. Disable/rotate all keys." A 15:12 update said multiple users had confirmed their AI keys were used. It also said Postgres and MongoDB secret rotation was now enabled for all users.

The company's last update that day, at 23:38 UTC, said: "Users have reported unauthorised use of AI API keys supplied to the Northflank platform." It told customers to revoke credentials with each provider, issue new ones and review AI provider usage and billing logs.

What Northflank customers should rotate

Northflank listed workload secrets, secret groups and global secrets as areas to rotate. It also advised turning off public access to database add-ons where it was enabled.

The UK security outlet teiss reported that Northflank also turned on audit logs for all customers, including free accounts. The company did not respond to teiss's questions about the attack route, the number of affected customers, or whether other credentials were abused.

Wiz 'Black October' report on AI developer platforms

Wiz's report, "Black October for AI Developer Platforms," published Oct. 30 by researchers Benjamin Read and Rami McCarthy, lists nine affected platforms: Fireworks AI, Andromeda, Zeabur, Lightning AI, OpenCode, Railway, Warp, CommandCode and Northflank. "Any other month, and this would be the top security story," the authors wrote.

The report says initial access varied. Wiz cites a supply chain compromise at Lightning AI, privilege escalation in build infrastructure at Railway, a third-party vendor at Warp and a "legacy system" at OpenCode. Attackers hunted for stored environment variables, especially AI access credentials.

Wiz analyzed two stolen enterprise keys. One served about 800,000 Bedrock calls over roughly 30 hours. The other served about 1.7 million calls over about 13 hours, more than 1 million of them successful. Wiz estimated combined list-price usage was likely above $100,000. First abuse came within 30 minutes of credential access.

Context: why AI keys are the prize

The practice of running models on someone else's account with stolen credentials is known as LLMjacking. Wiz said the same keys were used by several actors, which "suggests a robust market for this type of credential." It linked some infrastructure to discount model resale sites. Wiz also cautioned that those indicators are not directly tied to the Black October attacks.

Platforms like Northflank hold secrets for many customers in one place, so a single breach can expose keys for thousands of downstream services.

What's next

Northflank said its investigation was ongoing and that it had no further confirmed findings as of its 19:28 UTC update. Customers are waiting for a root-cause report that explains how decryption was possible and who was affected.

FAQ

Do I need to rotate keys if I use Northflank?

Yes. Northflank told all users to treat stored environment variables as compromised and rotate every secret, including AI provider keys.

How can I tell whether my AI key was abused?

Check usage and billing logs with each AI provider and report anything unusual to Northflank support and the provider, as the company advises.

Is this linked to the Railway incident?

Wiz lists both among nine October incidents with a similar focus on AI credentials, but it has not attributed them to a single attacker.

References

• Northflank Status — Northflank (Oct. 30, 2026)
• "Black October" for AI Developer Platforms — Wiz (Oct. 30, 2026)
• Northflank security incident leads to unauthorized use of AI API keys; customers told to rotate credentials — teiss (Nov. 2, 2026)