🔍 Read the full analysis: Fireworks AI Breach Widens To AI Assets: What To Check on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Fireworks AI’s November 9 update says its October security incident reached account and billing information and, for a limited set of accounts, secrets, datasets and incomplete models. Notion says chat transcripts from its AI Insights feature may have been in an accessed dataset, affecting fewer than 1% of workspaces. The number of affected customers, the cloud provider involved and what happened to accessed data remain unknown.
Fireworks AI says an October security incident reached further into customer information than the company initially disclosed: an update published November 9 says attackers accessed account and billing information and, for a limited set of accounts, secrets, datasets and incomplete models. Notion separately says transcripts from its Notion AI Insights feature may have been included in a dataset attackers could access, affecting fewer than 1% of workspaces, according to the company.
Fireworks’ account of the incident describes two access routes. In its October 9 bulletin, the company said someone had used internal Fireworks credentials to read environment variables from some customers’ evaluation jobs. Environment variables can contain configuration information, including API keys or tokens. Fireworks initially said it had not detected unauthorized access to customer models, datasets, training data, inference traffic or personal information.
The later update adds a separate event involving a third-party cloud provider that hosted Fireworks research and development machines. Fireworks says the provider experienced a security incident on October 23. On October 24, an unauthorized party used that access to reach Fireworks development machines and run an unauthorized workload, leading to additional customer data impact. The company says it detected the activity the same day and has not detected further successful access by the threat actor to the platform customers use since October 25.
Fireworks lists exposed information as resource metadata, billing information and, for certain customers, a subset of incomplete model weights and datasets. Billing information may include billing and shipping addresses, invoice and payment history, account balances and credit usage. The company says it does not store payment card numbers and that card numbers were not affected. It also says it has not detected impact to its Inference, Training SDK or Nexus products, which it describes as zero data retention by default.
Fireworks AI breach widens what customers should check
A November update says an October incident reached customer account and billing information, and limited accounts’ secrets, datasets, and incomplete models. Notion says AI Insights transcripts may have been in an accessed dataset.
What Fireworks says was accessed
The disclosed scope includes account and billing information; some accounts also had secrets, datasets, and incomplete models affected.
Resource and billing details
Billing information may include addresses, invoice and payment history, account balances, and credit usage.
Secrets and AI assets
Fireworks says some accounts’ secrets, datasets, and incomplete model weights were affected. Which secrets and how data was used remain unclear.
AI Insights transcripts
Notion says transcripts may have been in a dataset attackers could access; the share refers to workspaces, not users or transcripts.
How the scope expanded
Fireworks describes an initial access route and a later incident involving a third-party cloud provider.
What customers can review
Use the disclosed categories to guide a precautionary account review.
- Review account activity and relevant access logs.
- Check credentials and secrets associated with Fireworks.
- Identify datasets and incomplete models stored with the service.
- Monitor direct notices from Fireworks and updates from connected services such as Notion.
Risks to Customer AI Assets
The expanded scope matters because Fireworks provides infrastructure companies use to run and fine-tune open AI models. Customer datasets and model weights can contain proprietary business information or represent work invested in adapting a model for a specific purpose. Access to account secrets may also warrant review, although the update does not specify which secrets were accessed or whether any were used.
The Notion disclosure illustrates how an infrastructure incident can affect customers of companies that rely on the provider. Notion says chat transcripts from its AI Insights feature were in a dataset that attackers may have accessed and puts the proportion of affected workspaces at below 1%. That is a statement about the share of workspaces, not a count of users or transcripts. The number of affected Fireworks customers overall has not been disclosed.
Fireworks says it disabled self-service dataset uploads and downloads used by Batch Inference and dataset-based training while it investigates. That precaution may affect customers’ workflows, while the company’s statements that certain products were not affected and card numbers were not stored define what it says is outside the known impact. The investigation remains ongoing, so the current scope could change.
Popular right now
Roborock RockNeo Q110H Wire-Free Robot Lawn Mower
Apple AirPods Pro (1st Generation) with MagSafe Charging Case
Dyson Airsmooth Styling Brush
As an affiliate, we earn on qualifying purchases.
From October Alert to Expanded Scope
Fireworks’ first public bulletin, dated October 9, said the company had identified activity on October 6 and begun containment. It said it started notifying affected customers on October 7. The company later reported rotating employee API keys on October 9 and revoking more user API keys and rotating service-account tokens on October 10.
The cloud-provider incident adds a later stage to the timeline. Fireworks says the provider experienced a security incident on October 23 and the unauthorized workload ran the next day. Notion’s advisory describes receiving information from Fireworks between November 2 and November 8: an initial dataset was described as synthetic test data, then Fireworks told Notion that potential customer data may have been accessed, and the dataset was confirmed on November 8. Fireworks published its broader update on November 9.
Fireworks says it notified law enforcement and engaged Palo Alto Networks Unit 42 for forensic work. The dates and scope described here come from Fireworks’ security bulletins and Notion’s advisory; the provider’s identity and an independent account of the intrusion have not been provided in the source material.
“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 Use Still Unknown
Fireworks has not identified the third-party cloud provider or disclosed how many customers were affected. Its update also does not explain how the internal credentials used in the first access route were obtained, which specific secrets were accessed, or whether attackers copied, retained or used the data.
The company calls the model weights incomplete but has not explained whether that means partial copies, unfinished training runs or another condition. The available information does not establish whether the accessed datasets contained personal information beyond the Notion transcripts it identified, or whether any exposed material has been published or used for further attacks. Fireworks says its investigation is ongoing, and its statement that it has detected no further successful access since October 25 is limited to the evidence available to the company.
Customer Reviews and Investigation Updates
Fireworks says it is continuing its investigation with forensic support from Palo Alto Networks Unit 42 and has notified law enforcement. The company has not given a date for a final report or said when dataset uploads and downloads disabled during the investigation will resume. Customers may receive further notices as the investigation identifies affected accounts or clarifies the data involved.
Organizations that used Fireworks can compare their account activity and stored materials against the categories the company has disclosed, review relevant credentials and access logs, and follow any direct notice from Fireworks. Customers of services that may have placed data in affected datasets, including Notion’s reported AI Insights transcripts, should also monitor their provider’s security updates. These are precautionary steps; the public information does not establish that every customer or every type of data was affected.
Key Questions
What 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. The company also lists resource metadata among the information accessible through an affected internal application.
Were payment card numbers exposed?
Fireworks says it does not store payment card numbers and that they were not impacted. The company lists other billing details, including addresses, invoice and payment history, account balances and credit usage, as affected information.
What did Notion disclose?
Notion says transcripts from its AI Insights feature were in a dataset attackers may have accessed. It says fewer than 1% of workspaces were affected; it has not provided a total number of affected workspaces in the supplied information.
How many Fireworks customers were affected?
Fireworks has not disclosed a total. Its update refers to a limited set of accounts for some data categories, but that description does not provide a customer count.
What remains under investigation?
The cloud provider’s identity, the number of affected customers, how the internal credentials were obtained and what happened to accessed data remain unclear. Fireworks says the investigation is ongoing and may produce additional information.
Source: ThorstenMeyerAI.com
