Fireworks AI
Category: AI Infrastructure
Enterprise-grade platform for hosting, fine-tuning, and serving large language models (LLMs) and image models with high performance and cost efficiency. Fireworks AI was founded in 2022. The company is led by Lin Qiao. Based in Redwood City, California, USA. Team size: 51-200. Total funding raised: $327 million. Latest round: Series C, $250.0M, 2025-10, led by Lightspeed Venture Partners, Index Ventures, Evantic. Key investors include Sequoia Capital, Benchmark, NVIDIA, AMD, MongoDB Ventures, Databricks Ventures, Frank Slootman, Sheryl Sandberg, Lightspeed Venture Partners, Index Ventures, Evantic.
AMW Analysis
Fireworks AI provides enterprise infrastructure for hosting, fine-tuning, and serving large language and image models, with an emphasis on high-performance, lower-cost inference and flexible deployments including dedicated private clusters. Founded in 2022 and based in Redwood City, it has reported total funding of $327 million, most recently a $250.0 million Series C in October 2025.
Recent news points to accelerating demand for open-weight model deployment. In April, the company reported handling 15 trillion AI tokens per day, up from 10 trillion in late 2025, while highlighting GPU availability, hardware costs, and power-grid capacity as infrastructure constraints. By late May, reports said it was discussing a new round at an approximately $15 billion valuation. In July, Fireworks AI announced a $1.5 billion raise at a $17.5 billion valuation and annualized revenue exceeding $1 billion, alongside a fivefold year-over-year increase.
AMW analysis, generated from 4 tracked news signals.
- Founded
- 2022
- Headquarters
- Redwood City, California, USA
- Team size
- 51-200
- Total funding
- $327 million
Value proposition
Delivering the fastest inference speeds for open-source LLMs at lower costs, with flexible deployment options including dedicated private clusters for security and privacy.
Products and solutions
Hosted APIs for open-source LLMs (e.g., Llama, Mixtral, etc.), Serverless Fine-Tuning Services (LoRA), FireFunction V2: an open weights function calling model, Dedicated Private Cloud Deployments, Enterprise-grade SLAs and security features
Unique value
A proprietary inference engine optimized for ultra-low latency, a strong focus on the open-source model ecosystem, and the option for dedicated GPU clusters per customer for enhanced privacy and security.
Target customer
Enterprises, developers, and startups building generative AI applications that require scalable, fast, and cost-effective LLM inference.
Industries served
Technology/Software, E-commerce & Marketplaces, Customer Service, Legal Tech, Generative AI Applications
Technology advantage
Its proprietary inference engine delivers industry-leading speed for LLMs. A focus on open-source models prevents vendor lock-in for customers. Dedicated clusters offer a key advantage for enterprises with strict security and data privacy requirements.
How they differentiate
Differentiates through a focus on open-source models, superior inference speed at a lower cost, and flexible deployment options like dedicated clusters, contrasting with competitors' often proprietary, slower, or less flexible solutions.
Main competitors
OpenAI API/Azure OpenAI Service, AWS Bedrock, Together AI, Google Vertex AI
Key partnerships
Meta (PyTorch ecosystem), Hugging Face, LangChain, NVIDIA, Amazon Web Services (AWS), Google Cloud, Oracle, Microsoft Foundry (Azure)
Notable customers
Quora, DoorDash, Upwork, Cresta, Cursor, Liner, Superhuman, Sourcegraph, Tome, Samsung, Uber, Notion, Shopify
Major milestones
Raised $52M Series B in July 2024, led by Sequoia Capital, reaching a $552M valuation., Raised $25M Series A led by Benchmark in March 2024., Introduced FireFunction V2, an advanced function calling model., Demonstrated platform impact by tripling the response speed of Quora's Poe chatbot., Raised $250M Series C in October 2025 at a $4B valuation, co-led by Lightspeed, Index Ventures, and Evantic., Launched on Microsoft Foundry (Azure) in March 2026, bringing open-model inference to Azure., Scaled to 15 trillion tokens processed daily and $315M+ annualized revenue by early 2026.
Growth metrics
Grew to over 10,000 companies and hundreds of thousands of developers, processing over 10 trillion tokens daily as of October 2025, scaling to 15 trillion tokens daily by early 2026. Annualized revenue surpassed $280M as of October 2025.
Market positioning
Positioned as a leading enterprise provider for fast, scalable, and cost-effective open-source LLM hosting, fine-tuning, and inference.
Geographic focus
Primarily North America and Europe, but accessible globally via cloud APIs.
Patents and IP
No public patents listed; its proprietary inference engine (including custom kernels like FireAttention) is the core IP.
About Lin Qiao
Lin Qiao is the CEO and Co-Founder of Fireworks AI. She has a strong background in software engineering and AI, with previous leadership roles at major tech companies. Before founding Fireworks AI, she was a Senior Director of Engineering at Meta, where she led the PyTorch team. Her experience also includes positions at LinkedIn, Teradata, and Cisco Systems. She holds a Ph.D. in Computer Science from UC Santa Barbara.
Latest news about Fireworks AI
- Fireworks AI hits $17.5B valuation, $1B ARR as enterprises shift to cheaper open-weight models
- Fireworks AI in talks for $15B valuation, riding inference token boom to unicorn status
- Fireworks AI reportedly raising new round at $15B valuation, riding inference token boom
- Fireworks AI is handling 15 trillion AI tokens per day, up from 10 trillion in late‑2025, highlighting a massive surge in enterprise AI use. CEO Lin Qiao warns that GPU shortages, rising hardware cost
More AI Infrastructure companies
Official website: https://fireworks.ai