Fireworks AI closed a $1.5B Series D at a $17.5B valuation with NVIDIA participating — crossing $1B ARR as daily token volume jumps from 15T to 40T+.
Fireworks AI closed a $1.5 billion Series D at a $17.5 billion valuation, led by Atreides Management, Index Ventures, and TCV, with NVIDIA and more than a dozen other backers participating. The round follows a trajectory from $552M (Series B, July 2024) to $4B (Series C, October 2025) to $17.5B in 24 months.
Founder Lin Qiao — who led 300+ engineers building PyTorch at Meta — reports Fireworks crossed $1 billion in annualized revenue, up 5x year-over-year, with daily token volume growing from 15 trillion to more than 40 trillion. Customers include Samsung, GitLab, and enterprise AI teams running production inference at scale.
The deal lands in the same fortnight as Baseten's $1.5B Series F and Together AI's $800M Series C, confirming that inference infrastructure — not training — is where capital concentrates after the frontier-model megarounds.
Our Take
inference is a utilization game: winners sweat heterogeneous GPU fleets against bursty demand, not raw benchmark speed. Fireworks' PyTorch pedigree gives it credibility with developers who distrust hyperscaler lock-in — a narrative that supports premium valuation even as OpenAI, Anthropic, and Meta compress serving margins. For corp-dev, this category is still tuck-in ripe until a hyperscaler moves to acquire the remaining independent hosts.