🔍 Read the full analysis: Fireworks AI Breach: A Customer Checklist For Datasets And Models on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
Fireworks AI’s November 9 update says an incident first disclosed in October affected 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 and what happened to accessed data remain unclear.
Fireworks AI said in a November 9 update that an October security incident reached further than the company initially reported, exposing account and billing information and, for a limited set of accounts, secrets, datasets and incomplete models. Notion separately said transcripts from its Notion AI Insights feature may have been included in a dataset attackers could access, affecting fewer than 1% of its workspaces.
Fireworks, which provides services for running and fine-tuning open AI models, first disclosed suspicious activity on October 9. It said someone had used internal Fireworks credentials to read environment variables from customer evaluation jobs. In its November update, the company added that the incident affected account and billing information, and that data associated with some accounts included secrets, datasets and incomplete model weights.
The company also described a second route into its systems. A third-party cloud provider hosting Fireworks research and development machines experienced a security incident on October 23. Fireworks said an unauthorized party used that access the next day to reach development machines and run an unauthorized workload that caused additional customer data impact. Fireworks said it detected the activity on October 24.
Fireworks lists potentially accessed billing details such as billing and shipping addresses, invoice and payment history, account balances and credit usage. It says payment card numbers were not affected because it does not store them. The company also reported no detected impact to its Inference, Training SDK and Nexus products, which it describes as zero-data-retention by default. As a precaution, it disabled self-service dataset uploads and downloads used by Batch Inference and dataset-based training while its investigation continues.
Fireworks AI Breach Customer checklist
Fireworks AI’s November update expanded the reported scope of an October incident. Customers should review affected workflows and follow direct notices as the investigation continues.
What the update says
The listed data types do not mean every category was accessed for every customer.
Account and billing details
Potentially accessed details include addresses, invoices, payment history, balances and credit usage. Fireworks says it does not store payment card numbers.
Secrets, datasets and models
For a limited set of accounts, Fireworks reported impact to secrets, datasets and incomplete model weights.
AI Insights transcripts
Notion said transcripts may have been in an accessed dataset; its reported scope was fewer than 1% of workspaces.
Incident timeline
Fireworks’ account broadened after a second route through a third-party cloud provider.
Customer checklist
Use direct notices and review what your AI workflows contained.
- Check Fireworks’ direct notice to determine whether your account or assets were affected.
- Review evaluation jobs and dataset-based training for secrets or sensitive records.
- Follow incident communications for updated scope, credential guidance and restoration of dataset uploads and downloads.
- If you use Notion AI Insights, consult Notion’s advisory for your workspace assessment.
Risks for AI Customer Data
The expanded disclosure matters because companies that use hosted AI services may place business data, credentials and fine-tuning material with a provider, rather than handling all model operations on their own systems. Access to a dataset or secret can create risks beyond exposure of ordinary account details. Fireworks has not said that every listed category was accessed for every customer; it describes dataset and model access for a limited set of accounts.
Notion’s account gives the incident a concrete customer example. The company said transcripts from Notion AI Insights may have been in a dataset that attackers could access and that fewer than 1% of workspaces were affected. That is Notion’s reported scope for its workspaces, not a total count of Fireworks customers or an independently established measure of all data exposure.
For customers, the practical concern is determining whether information sent to affected jobs or stored in affected datasets included credentials, sensitive records or valuable model work. Fireworks’ statement that some models were incomplete narrows the description but does not establish whether the data could be used, reconstructed or otherwise exploited.
Products Worth Considering
Apple AirPods Pro (1st Generation) with MagSafe Charging Case
Bose QuietComfort Wireless Noise-Cancelling Over-Ear Headphones, Twilight Blue
Samsung Galaxy Tab A11+ 11-Inch Tablet, 6GB RAM, 128GB Storage, Expandable, Gray
As an affiliate, we earn on qualifying purchases.
From October Notice to Broader Update
Fireworks’ October 9 bulletin said the company had identified activity on October 6 and begun 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. At that stage, Fireworks said it had not detected unauthorized access to customer models, datasets, training data, inference traffic or personal information.
The later cloud-provider incident changed the picture. Fireworks says the provider experienced a security incident on October 23, and an unauthorized party reached Fireworks development machines on October 24. Fireworks says it detected that activity the same day and has not detected further evidence of successful unauthorized access on the platform customers use since October 25.
Notion said Fireworks contacted it on November 2 about a dataset initially described as synthetic test data. On November 6, Fireworks told Notion that malicious actors may have accessed potential customer data, and it confirmed the dataset on November 8. Fireworks says it notified law enforcement and engaged Palo Alto Networks Unit 42 for 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 Use Remain Unknown
Fireworks has not named the third-party cloud provider or said how the attackers first obtained the internal credentials involved in the October activity. It has not published a total number of affected customers or accounts, nor detailed which customers’ secrets, datasets or incomplete model weights were accessed. Its description does not explain what “incomplete” means for the model weights.
It is also unclear whether accessed information was copied, retained, shared or used after the intrusions. Fireworks has not publicly identified the attackers; the source material says the company cited evidence of links to a cybercriminal group but does not provide the group’s name or supporting details. The investigation remains ongoing, so the scope and attribution are not settled.
Customer Checks and Investigation Updates
Fireworks says it is continuing its investigation with Unit 42 and has notified law enforcement. Customers should use direct notices from Fireworks to determine whether their accounts, datasets or model assets were among those affected; the public update does not provide a complete customer list. Organizations that used evaluation jobs or dataset-based training can review what information those workflows contained and whether secrets or sensitive records were included.
Customers should also follow Fireworks’ incident communications for any updated scope, restoration of disabled dataset-upload and download functions, or further guidance on credentials and affected assets. Notion users can refer to Notion’s advisory for information about its own workspace assessment. The next major development will be a clearer account of the affected accounts and data, as well as any findings about how the attackers gained access and what they did with the information.
Key Questions
What did Fireworks AI say was affected?
Fireworks said account and billing information was affected and that, for a limited set of accounts, the incident reached secrets, datasets and incomplete models. It also listed resource metadata among information accessible through the affected internal application.
Were payment card numbers exposed?
Fireworks said it does not store payment card numbers and that they were not affected. It did say some billing information, including addresses, invoices, account balances and credit usage, was impacted.
Were Notion AI transcripts involved?
Notion said chat transcripts from its Notion AI Insights feature were in a dataset attackers may have accessed. Notion reported that fewer than 1% of its workspaces were affected.
How many Fireworks customers were affected?
Fireworks has not disclosed a total number of affected customers or accounts. The company describes access to secrets, datasets and incomplete models as affecting a limited set of accounts.
What should customers do now?
Customers should review notices from Fireworks, check whether affected workflows contained sensitive data or credentials, and follow any provider guidance about account-specific exposure. The company’s investigation is still underway, and its public update does not identify every affected account.
Source: ThorstenMeyerAI.com
