Was Fireworks AI inference traffic exposed in the breach?

Fireworks AI says it has not detected any impact to Inference, the Training SDK or Nexus, which run with zero data retention by default. The 9 November update stresses that the investigation is still ongoing. The confirmed impact is on resource metadata, billing information and, for a limited set of accounts, stored secrets, datasets and incomplete model weights.

Were credit card numbers stolen in the Fireworks AI hack?

No. Fireworks AI states that it does not store payment card numbers, so they were not affected. The billing data the attacker reached includes billing and shipping addresses, invoice and payment history, account balance and credit usage, according to the company's 9 November 2026 security bulletin update.

How did attackers get into Fireworks AI the second time?

Fireworks AI says a third-party cloud provider that hosted its R&D machines had a security incident on 23 October 2026. On 24 October the attacker used that foothold to reach Fireworks development machines and run an unauthorized workload. Fireworks detected it the same day and reported no further successful access to the customer platform after 25 October.

What has Fireworks AI changed after the breach?

Fireworks AI replaced credentials and service accounts, removed unneeded elevated privileges, isolated affected systems, tightened access to data storage and analytics, and hardened logging. Self-serve dataset uploads and downloads for Batch Inference and dataset-based training are disabled during the investigation, while the Training SDK stays available under zero data retention.

Fireworks AI · 2026-11-09 · major

Fireworks AI breach widens — customer datasets and model weights accessed

In short

Fireworks AI says attackers reached its development machines through a hacked cloud provider on October 24. They accessed billing data for customers and, for a limited set of accounts, secrets, datasets and incomplete model weights.

Fireworks AI · 2026-11-09 · 2 sources checked

Fireworks AI's October breach turned out bigger: billing data, and for some customers, datasets and unfinished fine-tuned weights.

Quick facts

Company	Fireworks AI (inference and fine-tuning platform)
First bulletin	9 October 2026 (stolen internal credentials)
Second intrusion	24 October 2026, via a third-party cloud provider
Data reached	Billing data; for some accounts secrets, datasets, incomplete model weights
Not affected (so far)	Inference, Training SDK, Nexus (zero data retention); card numbers not stored
Attribution	Financially motivated 'LLMJacking' group
Investigators	Palo Alto Networks Unit 42; law enforcement notified

What is it?

Fireworks AI published an update on 9 November 2026 to its October security incident. The first bulletin, on 9 October, said only secrets used in evaluation jobs had been reached. The company now says a second intrusion on 24 October reached account and billing information, and for a limited set of accounts also stored secrets, datasets and incomplete model weights.

How does it work?

According to the bulletin, a third-party cloud provider that hosted Fireworks' R&D machines was hacked on 23 October. The next day the attacker used that access to reach Fireworks development machines and run an unauthorized workload. Fireworks links the attack to a cybercriminal group doing 'LLMJacking' — stealing access to AI models and reselling the tokens through illicit routing gateways.

Why does it matter?

Teams that fine-tune on Fireworks may have had private training data and partly trained weights exposed, which matters more than a leaked API key that can simply be rotated. Fireworks has turned off self-serve dataset uploads and downloads for Batch Inference and dataset-based training during the investigation. Affected customers are being told directly what to do, so anyone with a notification email should act on it and rotate secrets stored on the platform.

Who is it for?

teams that fine-tune or serve models on Fireworks AI

Frequently asked questions

Was Fireworks AI inference traffic exposed in the breach?: Fireworks AI says it has not detected any impact to Inference, the Training SDK or Nexus, which run with zero data retention by default. The 9 November update stresses that the investigation is still ongoing. The confirmed impact is on resource metadata, billing information and, for a limited set of accounts, stored secrets, datasets and incomplete model weights.
Were credit card numbers stolen in the Fireworks AI hack?: No. Fireworks AI states that it does not store payment card numbers, so they were not affected. The billing data the attacker reached includes billing and shipping addresses, invoice and payment history, account balance and credit usage, according to the company's 9 November 2026 security bulletin update.
How did attackers get into Fireworks AI the second time?: Fireworks AI says a third-party cloud provider that hosted its R&D machines had a security incident on 23 October 2026. On 24 October the attacker used that foothold to reach Fireworks development machines and run an unauthorized workload. Fireworks detected it the same day and reported no further successful access to the customer platform after 25 October.
What has Fireworks AI changed after the breach?: Fireworks AI replaced credentials and service accounts, removed unneeded elevated privileges, isolated affected systems, tightened access to data storage and analytics, and hardened logging. Self-serve dataset uploads and downloads for Batch Inference and dataset-based training are disabled during the investigation, while the Training SDK stays available under zero data retention.