Fireworks AI Reports Security Breach Following Cloud Provider Hack

AI infrastructure startup Fireworks has reported a security breach that allowed attackers to access its development machines after breaching a third-party cloud provider.

Fireworks offers a high-performance inference platform for developers seeking to deploy and run open-source artificial intelligence models like Mistral, Meta’s Llama, or Google’s Gemma without the complexities of hardware limitations and infrastructure management.

The company informed law enforcement and its customers of the incident in early October, revealing that attackers accessed environment variables containing API keys and tokens stored as secrets for a limited number of accounts.

According to Fireworks, “The evidence indicates links to a cybercriminal threat actor group, which is financially motivated and aimed at selling tokens from unauthorized access to Large Language Models (LLMs) to illicit routing gateways, known as ‘LLMJacking.'”

After revoking compromised internal credentials, Fireworks discovered another security incident on October 23. In this new breach, the attackers accessed resource metadata, billing information, and incomplete model weights and datasets for some customers.

However, Fireworks stated that the attackers could not access customer payment card numbers, as the company does not store them.

Fireworks noted, “A third-party cloud provider that hosted Fireworks R&D machines experienced a security incident on October 23. On October 24, the unauthorized third party used it to reach Fireworks development machines and then run an unauthorized workload that led to additional customer data impact.”

The company detected this activity on the same day and initiated its incident response processes to remediate the situation and close known critical vulnerabilities. As of October 25, Fireworks has not detected further evidence of successful unauthorized access by the threat actor on the Fireworks platform.

In response to the attack, Fireworks has revoked and replaced credentials and service accounts, restricted unauthorized network pathways, isolated affected systems, and tightened access to data storage and analytics systems to prevent public or unauthorized access to stored data.

Fireworks is now collaborating with Palo Alto Networks’ Unit 42 incident response division to secure its systems against future attacks and investigate the full scope and impact of the incident.

Recently, researchers at Google’s Threat Intelligence Group (GTIG) and Mandiant reported that large language model hijacking attacks, also known as LLMjacking, are resulting in significant costs for compromised organizations, with cloud AI charges reaching as high as $1 million due to fraudulent calls using stolen API keys.

Source: BleepingComputer