Notion disclosed on November 11, 2026 that the breach at its AI infrastructure provider Fireworks AI exposed Notion AI chat transcripts from a small number of workspaces (fewer than 1% of its users), according to BleepingComputer.
What happened
- Fireworks AI had confirmed that an attacker accessed its development environment via a third-party cloud provider on October 23 and compromised resource metadata, billing information, model weights and datasets for some customers (see our earlier report on the second breach).
- Notion received an alert from Fireworks on November 2 about an exposed dataset that was said to contain only synthetic test data, cut off traffic to Fireworks, deleted its data there and halted training runs by November 3.
- On November 8, Fireworks confirmed that customer data had been accessed. The dataset included Notion AI chat transcripts used for the Insights feature, which has since been disabled.
- Notion says that for most affected workspaces the chats were "unattributable" (not labelled with author or workspace) and contained no tool call results.
Why it matters for privacy-focused users Fireworks had contractually committed to zero data retention for Notion data, yet transcripts were still in an exposed dataset. A "no retention" promise from a downstream inference provider is only as good as its enforcement. Services that route prompts to third-party hosts should be checked for which sub-processors see your chats.
Notion advises affected users to rotate passwords, API keys or other credentials entered into chats. Users who received no notice are most likely not affected.