Fireworks announced a $1.505 billion Series D on July 15 at a $17.5 billion valuation. Atreides Management, Index Ventures, and TCV led the round. Nvidia participated alongside Evantic Capital, Lightspeed, 20VC, Bessemer, Menlo, and others.
The company also published two operating figures: more than $1 billion in annualized revenue run rate and more than 40 trillion tokens served each day. Fireworks says over 95% of that volume comes from models specialized on customers' proprietary data and optimized for specific jobs. These are company disclosures; the announcement does not provide audited financials, gross margin, customer concentration, or retention figures.
| Fireworks disclosure | Figure |
|---|---|
| Series D | $1.505B |
| Valuation | $17.5B |
| Annualized revenue run rate | More than $1B |
| Daily token volume | More than 40T |
| Volume from specialized models | More than 95% |
Fireworks sits between model developers and application teams. Customers can bring open-weight or customized models while the platform handles serving infrastructure and optimization. The disclosed volume gives procurement teams a useful scale reference, though token count alone says little about latency, reliability, or cost for a particular workload.
Nvidia's participation gives the chipmaker financial exposure to a company aggregating inference demand across many customers. The announcement does not disclose Nvidia's investment size or any related supply agreement.
Teams evaluating Fireworks should ask four concrete questions: how the platform counts tokens, which regions and hardware serve each workload, who owns fine-tuned checkpoints, and how quickly an application can move to another provider. Those answers determine whether shared infrastructure improves a specific deployment or creates costly dependency.
The next useful disclosures would be gross margin, customer concentration, retention, and the split between base-model serving and customized workloads. Fireworks has put scale numbers on the table; those operating figures would show the quality of that scale.
The Signal is the public edge of a private practice. Sherpa points the same intelligence engine at one owner's business — competitors, suppliers, regulators, watched daily, graded and sourced. Work with a Sherpa →