AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: How To Check Your Exposure After The Fireworks AI Breach on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

  • Business-only prices and quantity discounts
  • Tax-exempt purchasing
  • Multiple users, one account, clear invoices
As an affiliate, we earn on qualifying purchases.

TL;DR

Fireworks AI’s November 9 update says an October security incident affected account and billing information and, for a limited set of accounts, secrets, datasets and incomplete models. Customers should review Fireworks’ notice, contact their security teams, and check for exposed credentials or data; the company has not disclosed how many customers were affected.

Fireworks AI said on November 9 that an October security incident affected account and billing information and, for a limited set of accounts, secrets, datasets and incomplete models. The updated disclosure broadens the company’s October 9 account, which had said it found no evidence that customer datasets, models or personal information were accessed. Customers using Fireworks should check for direct notices and review any credentials or data they entrusted to the platform.

Fireworks, which hosts and fine-tunes open AI models for companies, described two paths into its systems. In the first, detected October 6, someone used internal Fireworks credentials to read environment variables from customer evaluation jobs. Such variables can contain API keys or other credentials. The company said it began containment that day and started notifying affected customers October 7.

A separate incident involved a third-party cloud provider hosting Fireworks research and development machines. Fireworks said the provider experienced a security incident October 23, and an unauthorized party used access to reach Fireworks development machines and run an unauthorized workload on October 24. Fireworks said it detected that activity the same day. It has not named the cloud provider.

In its November 9 update, Fireworks listed resource metadata and billing information among the information accessible through the affected internal application. Billing details could include billing and shipping addresses, invoices, payment history, account balance and credit usage. For certain customers, the company said, a subset of datasets and incomplete model weights was accessed. It did not give a customer count or explain how much data was involved.

At a glance
updateWhen: Fireworks published the expanded disclo…
The developmentFireworks AI expanded its account of an October breach, saying attackers accessed account and billing information and some customers’ secrets, datasets and incomplete models.
Cybersecurity · Briefing

Check your exposure after the Fireworks AI breach

Fireworks AI’s November 9 update says the October incident reached account and billing information, and affected some customers’ secrets, datasets and incomplete models. Check your account status and act on any potentially exposed credentials or data.

What the update says

The expanded disclosure adds information beyond Fireworks’ October 9 bulletin.

1
Customer evaluation jobs

Internal credentials were used to read environment variables, which can contain API keys or other secrets.

2
Development machines

After a third-party cloud provider incident, an unauthorized workload ran on Fireworks development machines.

3
Expanded information

The update reports access to account and billing information and, for some customers, secrets, datasets and incomplete model weights.

What customers should check

Use Fireworks’ notice and your own records to determine whether your account or data may be involved.

  • Check email and account notifications for a direct Fireworks notice.
  • Compare the notice with evaluation jobs, dataset uploads and model training records.
  • Identify secrets that may have been present; revoke or rotate them as appropriate, and review access logs.
  • Review billing and account records for unexpected activity.
  • Assess whether potentially exposed data includes personal, confidential or regulated information.

What remains unclear

The company has not provided enough detail to establish each customer’s exposure.

Disclosed
  • Account and billing information was accessible.
  • Some customers’ datasets, secrets and incomplete models were affected.
  • Fireworks says it detected no further successful customer-platform access after October 25.
Not disclosed
  • How many Fireworks customers were affected.
  • Which accounts, or how much data, were involved.
  • Whether accessed data was copied or used.
  • The identity of the third-party cloud provider.
CaveatNotion said fewer than 1% of its workspaces were affected by a dataset that may have included Notion AI Insights transcripts. That figure describes Notion workspaces, not all Fireworks customers.

While the investigation continues

Fireworks says it is working with law enforcement and Palo Alto Networks Unit 42.

Preserve relevant logsFollow incident proceduresMonitor service updatesAsk support about your account
Key pointCheck your Fireworks account status directly; treat potentially exposed secrets as a security concern until you confirm their scope.
The takeaway
The expanded disclosure does not mean every customer was affected—use a direct notice, account records and security review to establish what needs action.

What Customers Should Check

The disclosure matters because Fireworks is an infrastructure provider: companies may send it training datasets, model files and credentials as part of their AI development work. Exposure at a service provider can create follow-up risks for customers even when their own systems were not directly breached. A leaked token, for example, may need to be treated as exposed until its owner confirms its scope and status.

Customers should start by checking email, account notifications and any message from Fireworks for a notice identifying the affected account or data. Security teams can compare the notice with records of evaluation jobs, dataset uploads and model training during the relevant period. If credentials or secrets could have been present in an affected environment, teams should identify them, revoke or rotate them as appropriate, and review access logs for use they do not recognize. Those are precautionary steps, not evidence that a particular customer’s credentials were stolen.

Organizations should also review billing and account records for unexpected changes or activity, and check whether any exposed dataset included personal, confidential or regulated information. If it did, the organization should assess its own reporting and contractual duties. Fireworks has not said that every customer’s data was affected, so customers should not assume either that they were exposed or that they were unaffected without checking their account status.

Products Worth Considering

Samsung Galaxy Tab S7 11-inch Android Tablet 128GB Wi‑Fi with S Pen Mystic Black

Samsung Galaxy Tab S7 11-inch Android Tablet 128GB Wi‑Fi with S Pen Mystic Black

Jackery Explorer 300 Portable Power Station with 40W Solar Panel

Jackery Explorer 300 Portable Power Station with 40W Solar Panel

As an affiliate, we earn on qualifying purchases.

From October Notice to Expanded Scope

Fireworks’ first public bulletin, dated October 9, described misuse of internal credentials and access to environment variables in a limited number of customer accounts. At that point, the company said it had not detected evidence of access to customer models, datasets, training data, inference traffic or personal information. The November 9 update added account and billing information and data from certain accounts, changing the stated scope.

The company said it rotated employee API keys October 9 and revoked more user API keys and rotated service-account tokens October 10. It also said it had not detected further successful unauthorized access by the actor on the customer platform after October 25. Fireworks said it notified law enforcement and engaged Palo Alto Networks Unit 42 for forensic work. Notion separately said Fireworks contacted it in early November about possible access to customer data.

Fireworks says it does not store payment card numbers, which it said were not affected. It also reported no detected impact to Inference, Training SDK and Nexus, products it describes as zero-data-retention by default. That statement concerns those named products; it does not negate the company’s separate report of access to datasets and other information through the affected systems.

“Since then, we discovered impacts to account and billing information.”

— Fireworks AI, November 9 security update

Exposure Details Still Missing

Fireworks has not disclosed how many customers were affected, which accounts were involved, or the amount and contents of the datasets and incomplete model weights accessed. It has also not identified the third-party cloud provider or explained how the internal credentials used in the earlier activity were first obtained. The company has not said whether the accessed data was copied, shared or otherwise used.

Notion said chat transcripts from its Notion AI Insights feature were in a dataset attackers may have accessed, and that fewer than 1% of workspaces were affected. That figure is Notion’s description of its own workspaces, not a count of all Fireworks customers. Notion has not established from the supplied information that every transcript in the dataset was read or taken.

Fireworks said it had not detected further successful access to the customer platform after October 25, but its investigation remains open. Its statements about what was not affected are based on what the company says it has detected so far; the available disclosure does not provide independent forensic findings or a full account of the attacker’s actions.

Customer Notices and Investigation

Fireworks says it has disabled self-service dataset uploads and downloads used by Batch Inference and dataset-based training while it investigates. The company has not provided a date for restoring those functions. Customers relying on them should monitor Fireworks’ service communications for changes and use the company’s support channel to ask whether their account, projects or datasets were included.

The next public milestones are likely to be further findings from Fireworks’ investigation and additional customer-specific notifications. Fireworks says it is working with law enforcement and Unit 42; it has not announced a final report or a completion date. Until it provides more detail, customers should preserve relevant logs, follow their incident-response procedures and treat potentially exposed secrets as a security concern where their account information warrants it.

Key Questions

How can I find out whether my Fireworks AI account was affected?

Check for a direct notice from Fireworks, review account communications and contact its support team if your status is unclear. The public update does not list affected customers or provide a count.

What information does Fireworks say may have been accessed?

Fireworks lists resource metadata and billing information, and says a limited set of accounts may also have been affected through access to secrets, datasets and incomplete model weights. The company has not detailed the contents or volume for each account.

Were payment card numbers exposed?

Fireworks says it does not store payment card numbers and that they were not affected. It separately says billing and payment-history information was accessible.

What should a customer do about API keys or other secrets?

If a Fireworks notice or internal review indicates a credential may have been in an affected environment, identify it and revoke or rotate it under your organization’s security procedures. Review relevant logs for activity you do not recognize; the disclosure does not establish that every customer’s secrets were accessed.

Is the investigation finished?

No. Fireworks says its investigation is ongoing and could reveal more information. The company has not named the cloud provider or disclosed the total number of affected customers.

Source: ThorstenMeyerAI.com

You May Also Like

Harnessing Anthropic Claude Apps Gateway For Enterprise AI On AWS

AWS has published guidance on deploying an Anthropic Claude apps gateway for enterprise workloads, but technical details and availability are still unconfirmed.

Understanding SenseTime’s AI Revenue Breakdown On HKEX

TradingView lists a revenue breakdown for SenseTime Class B shares on HKEX, but no financial figures or details have been confirmed or disclosed.

The Intersection Of Mathematics And AI: Insights From The Advisory Group

OpenAI has announced the creation of an advisory group focused on the intersection of mathematics and artificial intelligence, aiming to guide its AI research and applications.

12 AI Coding Tools For New Skills And Software Projects In 2027

A source-based guide to 12 AI coding books, from beginner introductions to lifecycle, governance, Claude Code and project-specific titles.