Startups

Fireworks AI hits $1B revenue, $17.5B valuation on open-source wave

The Nvidia-backed cloud startup grew 5x in a year as enterprises seek cheaper alternatives to frontier models from OpenAI and Anthropic.

Omega Editorial· July 16, 2026· 3 min read

Fireworks AI reaches $1 billion revenue milestone

Fireworks, a cloud infrastructure startup backed by Nvidia, announced Thursday it has surpassed $1 billion in annualized revenue—five times its level from a year ago. The company also closed a $1.5 billion funding round at a $17.5 billion valuation, according to details first reported by CNBC.

The San Mateo-based company competes with Amazon, Google, and Microsoft to host AI models that developers integrate into applications. Its rapid growth reflects mounting enterprise interest in open-source alternatives to expensive proprietary models from OpenAI and Anthropic.

Why it matters

Fireworks' trajectory demonstrates that the AI infrastructure market remains far from consolidated. While hyperscalers dominate cloud computing overall, specialized providers are capturing meaningful share by offering cost advantages and flexibility around model choice. For enterprises balancing AI ambitions against budget constraints, platforms that deliver "five to 10 times cheaper" inference—as CEO Lin Qiao claims—present a compelling value proposition that challenges the frontier labs' pricing power.

Cost pressures drive open-model adoption

Qiao told CNBC that finance executives are increasingly directing teams toward open-source models to manage AI spending. Coinbase CEO Brian Armstrong publicly noted in June that his company has been adopting cheaper models where appropriate.

"Our cost compared with the equivalent-quality closed model is five to 10 times cheaper," Qiao said in an interview at the company's headquarters.

Fireworks positions itself as enabling "specialized intelligence" rather than the "generalized intelligence" offered by frontier labs. The platform allows clients to bring proprietary data and refine open-weight models—including those from Chinese companies like DeepSeek, MiniMax, and Z.ai, as well as OpenAI's released models—for specific use cases.

Scale and partnerships

The startup now processes 40 trillion AI tokens daily, Qiao said. For context, Google disclosed in May that its AI models handled about 19 billion tokens per minute for developers (roughly 27 trillion per day), while OpenAI announced 15 billion tokens per minute in March (approximately 22 trillion daily).

Fireworks has diversified beyond its early concentration—about half its revenue came from AI coding startup Cursor as of last year. The client base now includes Elastic, GitLab, and MongoDB. In June, SpaceX agreed to acquire Cursor in a $60 billion stock deal expected to close this quarter.

In March, Fireworks announced a partnership with Microsoft that allows the tech giant's customers to access models through Fireworks' infrastructure, which draws computing power from more than 20 suppliers.

Expansion plans

Qiao, a former Meta director who co-founded Fireworks in 2022 with six colleagues, said the company employs around 200 people and expects to reach 600 by year-end 2026. The startup hired former Salesforce executive George Hu as president in April and plans to build out a sales organization after years of relying on developer self-service.

Atreides Management, Index Ventures, and TCV led the new funding round, with participation from Nvidia, Evantic, and Lightspeed Venture Partners.

CNBC first reported the funding details and revenue milestone.

#fireworks ai#ai infrastructure#open-source ai#nvidia#ai cloud#enterprise ai

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Startups

Startups· 3 min read

Andreessen Horowitz launches $1.1B fund for AI infrastructure

The Machine Age Fund will back startups building chips, data center equipment, edge devices and robotics as hardware struggles to keep pace with AI demand.

Via AI Watch · Aug 29, 2026
Startups· 3 min read

Nvidia, Stripe Spend $26B on Open-Weight AI Acquisitions

Three major deals in August signal tech giants are hedging against frontier labs by buying into the open model ecosystem.

Via AI Watch · Aug 28, 2026
Startups· 2 min read

Andreessen Horowitz Launches $1.1B Fund for AI Infrastructure

The venture firm's new Machine Age fund marks a strategic shift toward hardware investments supporting AI's physical buildout.

Via AI Watch · Aug 28, 2026