AI Stack Alert Medium TD-2026-0057 For CISO, SOC / IR, AI Eng, AppSec Unconfirmed AI-assisted
Fireworks AI breaches expose billing data, secrets and customer AI data; LLMJacking trend noted
AI infrastructure company Fireworks says attackers stole account and billing information and secrets during breaches discovered October 6 and October 23. Fireworks said a limited number of customer accounts were affected and that those customers were notified. Google's Mandiant and GTIG reported multiple LLMJacking incidents in 2026, in which attackers hijack enterprise access to large language models.
Why it matters
Organizations using Fireworks or similar inference providers may have exposed credentials, billing data and AI datasets. LLMJacking can generate fraudulent cloud charges of over $100,000 in a single day, according to GTIG.
Do this
- Identify any use of Fireworks APIs and rotate API keys, tokens and secrets that were stored with or shared with the provider.
- Revoke and reissue credentials for any Fireworks accounts, and review access logs for unexpected activity.
- Set cloud and LLM provider billing alerts for sudden spikes in consumption to detect LLMJacking early.
- Review third-party vendor notices and confirm whether downstream features that depend on Fireworks have been disabled or re-routed.
Signals
- NoCISA KEV
- 26/100Priority
- 70%Confidence
Evidence
“Fireworks published an updated notice on Monday saying it hired Palo Alto Networks Unit 42 to investigate a data breach it initially discovered on October 6.”
“Across individual incidents, we observed fraudulent charges associated with resource consumption exceeding $100,000 in a single day”
Sources
Affected / related
Fireworks, Palo Alto Networks Unit 42, Notion, Mandiant, LLMJacking
Published 2026-11-10 16:08 UTC · updated 2026-11-12 16:47 UTC · AI-assisted summary, checked against quoted sources · Unconfirmed: single non-official source · How we score