How much has Fireworks AI raised?

Fireworks AI has raised over $327 million in disclosed equity through October 2025, spanning a seed round led by Benchmark and three subsequent priced rounds — Series A ($25M, March 2024), Series B ($52M, July 2024), and Series C ($250M, October 2025). All rounds to date have been structured as equity, with no disclosed debt or venture debt facilities, and every priced round has been at a higher valuation than the prior one — no down rounds have occurred.

If the rumored Series D closes at the $15B valuation reported by Bloomberg in May 2026, total disclosed funding would likely exceed $500M depending on the size of that round. The compound growth rate of Fireworks' valuation since its $552M Series B (July 2024) to the rumored $15B Series D (mid-2026) would represent roughly a 27x increase in approximately 23 months — one of the steeper valuation trajectories in the AI infrastructure cohort and comparable in pace to the fastest-growing infrastructure companies of the 2020s.

Importantly, the valuation trajectory has been driven by genuine revenue growth, not purely multiple expansion. ARR grew from approximately $50–70M at Series B to $280M at Series C to $800M annualized by May 2026 — suggesting the per-dollar-of-revenue valuation multiple has actually compressed slightly even as the absolute valuation has soared.

Benchmark led both the seed round and Series A, reflecting early conviction in the PyTorch founding team's ability to commercialize Meta-scale infrastructure — a pattern consistent with Benchmark's portfolio of technical founder-led infrastructure companies. Sequoia Capital led the Series B in July 2024 and continued at Series C, underscoring sustained institutional conviction across two rounds and a 7x valuation step-up.

Strategic investors NVIDIA and AMD both joined the Series B — a particularly significant signal. Both GPU chipmakers backing the same inference platform indicates Fireworks' position as a neutral aggregator that drives silicon demand regardless of which GPU architecture wins. Lightspeed Venture Partners and Index Ventures co-led the $250M Series C alongside first-time investor Evantic. Index is also set to co-lead the Series D, making it the most consistent institutional backer across the late-stage rounds.

Angel investors from the Series A — Frank Slootman, Sheryl Sandberg, Howie Liu, and Alexandr Wang — represent some of the most credentialed enterprise software and AI operators in Silicon Valley, providing both capital and warm enterprise introductions to the company's earliest customer conversations.

The $4B Series C valuation (October 2025) was set when Fireworks had approximately $280M in annualized revenue, implying roughly a 14x ARR multiple. By May 2026 — just seven months later — ARR had reached approximately $800M, representing roughly 186% revenue growth within that window. The Series D is being sought at $15B, representing approximately a 19x forward ARR multiple and a 3.75x step-up in absolute valuation in under a year.

The driver is the compound AI and agentic AI wave. Large enterprises building coding assistants, autonomous agents, and multi-model pipelines consume orders of magnitude more tokens per workflow than single-model chatbot use cases — a dynamic that directly benefits per-token inference providers like Fireworks. As the 'inference per useful task' ratio rises, platforms that compete on cost-per-token and latency capture disproportionate share of expanding workloads.

Additionally, the competitive moat has widened. The December 2025 NVIDIA licensing deal with Groq effectively took the most compelling low-latency custom-silicon alternative off the independent market. Baseten and Together AI remain strong competitors but have not demonstrated Fireworks' combination of token volume growth, customer breadth, and proprietary kernel-level optimization. The Series D valuation premium reflects the market's view that Fireworks has established a durable position in a winner-takes-most infrastructure category.

Fireworks AI has not publicly reported profitability as of mid-2026. Gross margin was approximately 50% at Series C with a target of 60%, suggesting the business generates meaningful gross profit but is reinvesting heavily in headcount, global infrastructure expansion (a planned 3–4x compute scale-up was announced at Series C), and go-to-market build-out. At approximately 150 employees and $800M ARR, revenue per employee exceeds $5M annualized — an unusually high figure for a company at this stage and a strong indicator of eventual operating leverage as headcount grows more slowly than revenue.

No IPO timeline has been publicly stated. At the rumored $15B Series D valuation, a public market exit would likely require $1.5–2B+ in ARR at meaningful margins to attract institutional buyers at a 7–10x revenue multiple. If 2026 growth continues at even half the pace of the prior twelve months, Fireworks could cross the $1.5B ARR threshold by early 2027, putting a potential 2027–2028 IPO window within reach — particularly if agentic AI adoption continues to drive token volume and margin improvement.

A company at Fireworks' scale and growth trajectory typically enters a rapid infrastructure and tooling procurement cycle following each major funding round. The planned 3–4x compute infrastructure expansion announced at Series C (October 2025) directly signals budget for cloud infrastructure, GPU capacity contracts, networking, observability, and data infrastructure spend. If the Series D closes, a new wave of enterprise tooling procurement is likely, particularly in GTM, security and compliance, and developer tooling categories.

Fireworks' bottoms-up developer motion means the company skews toward self-serve vendor evaluation for tools under approximately $50K/year — technical champions inside engineering teams drive initial adoption, which then escalates to procurement. For six-figure-plus contracts, the primary economic buyer is typically a VP of Engineering or Head of Infrastructure, with the CTO providing technical sign-off on architecture decisions. Given the pace of hiring in sales and marketing (VP Sales, CMO both added 2024–2025, active AE recruiting in 2026), the GTM infrastructure itself is an active investment area.

Fireworks is listed on AWS Marketplace and GCP Marketplace, which means enterprise vendors with those marketplace listings can leverage existing cloud commit balances to accelerate purchasing. For sellers offering observability, security, data infrastructure, or ML tooling that integrates with PyTorch or Kubernetes, Fireworks represents a high-propensity target with genuine budget and a fast-moving procurement motion.