TL;DR
Get business pricing on networking and server gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
Fireworks AI’s November 9 update says attackers accessed account and billing information and, for a limited set of accounts, secrets, datasets and incomplete models. Notion says transcripts from its AI Insights feature may have been in an accessed dataset, affecting fewer than 1% of workspaces. The number of affected customers and what happened to the data remain unknown.
Fireworks AI said on November 9 that an October security incident reached further than it initially reported, with attackers accessing account and billing information and, for a limited set of accounts, secrets, datasets and incomplete models. The update matters to businesses that use the platform to run or fine-tune AI models because some customer material may have been exposed.
Fireworks’ October 9 bulletin said the company had detected misuse of internal credentials and access to environment variables in a limited number of customer accounts. At the time, it said it had not detected unauthorized access to customer models, datasets, training data, inference traffic or personal information. Its November update revised that account, saying it had since discovered impacts to account and billing information and, for certain accounts, secrets, datasets and incomplete models.
The company described two routes into its systems. In the first, someone used what Fireworks called “internal Fireworks credentials” to read environment variables from customer evaluation jobs. In the second, a security incident at a third-party cloud provider hosting Fireworks research and development machines was followed by an unauthorized workload on development machines on October 24. Fireworks said it detected that activity the same day.
Fireworks lists exposed information as resource metadata, billing details and, for certain customers, a subset of incomplete model weights and datasets. Billing details may include addresses, invoice and payment history, account balances and credit usage. The company says it does not store payment card numbers and has not detected impact to its Inference, Training SDK or Nexus products, which it says are zero data retention by default.
Risks for AI Platform Customers
The incident reaches beyond billing records because datasets, secrets and model weights can contain commercially sensitive material. A company may use a hosted AI platform to tune a model on proprietary data or store credentials that connect services. If such material was accessed, the exposure could create business, privacy or security concerns, though the available information does not establish that any particular dataset or model can be used by an attacker.
Notion said chat transcripts from its Notion AI Insights feature were included in a dataset that attackers may have accessed. It put the affected share at fewer than 1% of workspaces. That statement gives one concrete example of a downstream customer concern, while leaving the total scope across Fireworks customers unknown.
Fireworks says it has not detected further successful unauthorized access by the threat actor on the customer platform since October 25. That is a statement about what its investigation has found so far; it does not establish what happened to information accessed earlier or whether copies were retained elsewhere.
Products Worth Considering
Logitech MX Ergo S Advanced Wireless Trackball Mouse USB-C – Graphite
Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation
Bose Lifestyle Ultra Compact Wireless Smart Home Speaker Bundle, Black
As an affiliate, we earn on qualifying purchases.
From Initial Disclosure to Update
Fireworks first disclosed the incident on October 9, after identifying suspicious activity on October 6 and beginning containment. It said it started notifying affected customers on October 7, rotated employee API keys on October 9 and revoked additional user API keys and rotated service-account tokens on October 10.
The company’s November account added a second phase. It said the third-party provider experienced a security incident on October 23, and an unauthorized party used the access to reach Fireworks development machines and run a workload the following day. Fireworks said that activity caused additional customer data impact and was detected on October 24.
Notion said Fireworks contacted it on November 2 about a dataset initially identified as synthetic test data. On November 6, Fireworks told Notion that malicious actors may have accessed potential customer data, and the dataset was confirmed on November 8, according to Notion’s advisory. Fireworks says it notified law enforcement and engaged Palo Alto Networks Unit 42 to conduct forensic work.
“For a limited set of accounts, we also discovered impact to secrets, datasets, and incomplete models.”
— Fireworks AI, November 9 security update
Scope and Data Fate Remain Unknown
The public updates do not identify the third-party cloud provider, give a total number of affected Fireworks customers or explain how the internal credentials were first obtained. Fireworks also has not explained what “incomplete” means for the affected model weights, such as whether they were partial copies or unfinished training outputs.
It remains unclear whether attackers removed, retained, shared or used the accessed data. The statements that Fireworks has not detected further access and that certain products were not impacted describe the company’s findings to date; the investigation is ongoing and Fireworks says additional information may emerge.
Forensic Review and Customer Notices
Fireworks says its investigation with Palo Alto Networks Unit 42 is ongoing and that it has notified law enforcement. The company may provide further details as the forensic review clarifies which accounts and data were affected. Customers that received notices can use those communications to determine whether their own accounts or material are implicated.
Further updates from Fireworks and affected customers such as Notion may clarify the scale of the exposure and whether accessed information was used or distributed. Until then, the confirmed public account remains limited to the categories Fireworks has disclosed and the specific Notion workspace estimate.
Key Questions
What information does Fireworks AI say attackers accessed?
Fireworks says attackers accessed account and billing information and, for a limited set of accounts, secrets, datasets and incomplete models. It also lists resource metadata and says payment card numbers were not impacted because it does not store them.
Were model weights exposed?
Fireworks says a subset of incomplete model weights was accessed for certain customers. It has not publicly defined what “incomplete” means or described how the affected weights could be used.
Were Notion AI transcripts affected?
Notion said chat transcripts from its AI Insights feature were in a dataset attackers may have accessed. It said fewer than 1% of workspaces were affected; the advisory does not establish that attackers used or shared the transcripts.
Did Fireworks say its AI products were affected?
Fireworks said it has not detected impact to Inference, Training SDK or Nexus, which it describes as zero data retention by default. Its investigation is ongoing.
What is still unknown about the breach?
The company has not disclosed the cloud provider involved, the total number of affected customers, how the internal credentials were first obtained or what happened to the accessed data.
Source: Thorsten Meyer AI
