Fireworks Cyberattack Exposes Customer Data And AI Models As Fallout Widens

Fireworks, a popular infrastructure provider that helps large enterprises and startups build and run their own AI models, said Monday some of its customers’ datasets and models were accessed by hackers over the course of the past month.

The breaches expose a risk in corporate America’s rush to build cheaper, bespoke AI models: in doing so, they often send proprietary information, like sensitive customer data, to platforms like Fireworks. And a compromise at Fireworks would potentially expose information of its customers, making it an attractive target for hackers.

Fireworks counts roughly 10,000 enterprise custo­mers—including Uber, Shopify, Upwork and DoorDash as customers, as well as marquee startups like Cursor and Cognition—who use Fireworks to build specialized AI using their own proprietary data on top of cheaper, often Chinese models.

The company booked about $250 million in 2025 revenue, Forbes reported last month. Its valuation has grown fourfold to $17.5 billion in the past nine months. In July, CEO Lin Qiao became a new billionaire with an estimated 6% stake worth $1 billion.

Forbes spoke with two customers who had substantially moved their business off Fireworks’ systems. One described it as a series of “very bad hacks” of increasing frequency and criticized Fireworks’ customer communications around the attacks. Two other customers said they were not aware of any Fireworks hacks.

In a security update, the AI company said the cybersecurity incident was a result of a technique known as “LLMjacking,” where attackers steal access to AI models and resell it without permission, leaving the owner with the bill. In a statement shared with Forbes, the company said it is working with Unit 42, the cybersecurity division of Palo Alto Networks to investigate recent attacks.