AI inference giant Fireworks, with $1 billion in annualized revenue, hit by two consecutive hacks as major clients consider jumping ship

Beating AI News Flash: AI inference and model training platform Fireworks has suffered two consecutive hacker intrusions. On November 9, the company admitted that some customers' billing information, datasets, and unfinished model weights were accessed without authorization. Two customers interviewed by Forbes said they have moved most of their business away from Fireworks.

The first intrusion occurred in early October. Hackers stole Fireworks' internal credentials and obtained API keys that some customers used for model evaluation. At the time, the company said it had not found that customer models and datasets had been accessed.

On October 24, hackers again entered Fireworks' R&D machines through a compromised third-party cloud service provider, ran unauthorized programs, and further affected customer data.

One customer interviewed by Forbes complained that hacker attacks on Fireworks are becoming increasingly frequent and that communication with customers has been poor. Some major customers who pay Fireworks millions of dollars a year are considering changing suppliers.

Fireworks' annualized revenue had reached $1 billion in July this year, with a valuation of $17.5 billion. The company had originally planned to launch a new funding round soon, with a target valuation that could reach more than $30 billion. Customer attrition could put pressure on this financing.

Fireworks said the attackers may be linked to the black market that steals and resells large-model invocation quotas. So far, no impact has been found on core services such as online inference, and payment card numbers have not been leaked. The company has suspended some data upload and download functions and invited Unit 42 under cybersecurity company Palo Alto Networks to assist with the investigation.