Fireworks AI, the Redwood City startup that rents out computing infrastructure for running and customizing AI models, is in early talks to raise new funding at a valuation between $30 billion and $40 billion, Business Insider reported on October 20, citing people familiar with the matter. The target would roughly double the company’s value just three months after it closed a $1.5 billion Series D at a $17.5 billion valuation.

The discussions remain at an early stage, and people familiar with the matter cautioned that “the financing talks are still in early stages and that valuation targets could change.” Fireworks declined to comment. The new report follows The Information’s reporting last month that the company had already been weighing a round at a $30 billion valuation.

What Fireworks AI does

Founded in 2022 by former Meta and Google AI engineers, Fireworks sells what it calls “specialized intelligence”: leased GPU capacity plus the tooling for businesses to fine-tune and serve AI models built around their own proprietary data and workflows, rather than building a frontier model of its own. Chief executive Lin Qiao previously led Meta’s PyTorch team, where she managed more than 300 engineers building the open-source framework that still trains and runs a large share of the world’s AI models, Startup Fortune notes.

Data card showing Fireworks AI key figures: $30-40B valuation in early talks, $17.5B Series D, $1B+ annualized revenue, 40 trillion tokens per day

Graphic: GPTOnline, using Fireworks assets

The company says it now processes more than 40 trillion tokens a day, up from roughly 15 trillion a year ago, and has crossed $1 billion in annualized revenue, about five times what it was generating last year. More than 95% of that traffic comes from models customers have fine-tuned on their own proprietary data, not off-the-shelf open weights, a sign of how strongly enterprises are adopting specialized AI systems rather than general-purpose chatbots.

A steep climb in nine months

Fireworks announced its $1.5 billion Series D in July at a $17.5 billion valuation, led by Atreides Management, Index Ventures and TCV, with participation from Nvidia, Lightspeed Venture Partners, Bessemer Venture Partners and Menlo Ventures. That followed a $250 million Series C at a $4 billion valuation in late 2025. A raise at $30 billion to $40 billion would value the four-year-old company at roughly ten times its late-2025 mark.

Fireworks operates in a crowded layer of the AI market, competing with startups such as Together AI, Baseten and Replicate, while also facing pressure from large cloud providers including Amazon Web Services, Google Cloud, Microsoft Azure and Oracle. The new fundraising effort signals that investors still see room for infrastructure specialists despite the dominance of major cloud platforms.

Security incident in the background

The talks come shortly after the company disclosed a security incident earlier this month. Fireworks said unauthorized use of internal credentials allowed an intruder to access data from some customer accounts, and Qiao said the company would be transparent about what happened and what it finds, Traders Union reports. How the company handles the fallout bears directly on the enterprise trust its business depends on.

Why it matters

A $30 billion to $40 billion valuation would test how much investors will pay for AI infrastructure built around customer-specific models at a moment when the inference layer, the business of actually serving models to users, is consolidating fast. Fireworks’ numbers, $1 billion-plus in annualized revenue on 40 trillion daily tokens, give investors something most AI startups cannot offer: a real, fast-growing revenue base. If the round lands near the top of the reported range, it would cement the company as one of the most valuable private AI infrastructure firms in the world, and raise the bar for rivals racing to serve the fine-tuning economy.

Editorial graphic: Fireworks AI funding round in early talks could double valuation to $30-$40B

Graphic: GPTOnline, using Fireworks assets

Sources