Fireworks AI confirms LLMJacking attack hit keys, billing data and partial model weights
FinanceYF5 · x · 2026-11-12
Fireworks AI's updated security bulletin confirms its October incident was broader than first disclosed: account/billing info was affected, and for some customers secrets, datasets and incomplete model weights were accessed. Evidence points to a financially motivated cybercriminal group selling LLM access to illicit routing gateways ("LLMJacking"). Attackers reached dev machines via a breached third-party cloud provider on Oct 24; remediation followed the same day, with no new successful access detected after Oct 25. Palo Alto Networks Unit 42 assisted forensics; payment card numbers were not stored or impacted.
Related event: Fireworks AI confirms breach exposing keys and model weights(2 posts)→
More from Safety
- SoftBank found 15,000 vulns with OpenAI's tool, then GPT-6 Astra surfaced 35,000 more — IgorCarron · 2026-11-12
- Paul Christiano's three-night Hinton Lectures on governing transformative AI earn rave reviews — S_OhEigeartaigh · 2026-11-12
- Redwood CEO: Leaked algorithmic secrets alone could wipe out the US AI lead over China — connoraxiotes · 2026-11-12
- LLM audit-log eval: 0/40 lies unprompted, 101/101 fake a check when asked — mrsergeiluzhnov · 2026-11-12
- Viral AI experiment fills an empty room with everything platforms know about you — nikola_mr64990 · 2026-11-12
- Practitioner breaks down zero-data-retention terms: Anthropic defaults to 30 days, OpenAI needs account-team negotiation — HanchungLee · 2026-11-12