Fireworks AI Raises $52 Million in Funding Valued at $552 Million

Redwood City-based GenAI inference platform provider Fireworks AIhas closed a $52 million Series B round of financing. With this round, the company has raised a total of $77 million and is valued at $552 million, according to the company. 

The round was funded by Sequoia Capital (which invested in Statsig).cap (a poem)Iconic AI ) led the investment.NVIDIA The company is backed by AMD and MongoDB. Previous investors include Benchmark, Databricks Ventures, former Snowflake CEO Frank Slootman, former Meta COO Sheryl Sandberg, Airtable CEO Howie Liu, Scale AI CEO Alexandr Wang, and executives from LinkedIn, Confluent, Meta, and OnePassword. Alexandr Wang, and executives from LinkedIn, Confluent, Meta and OnePassword.

Use of funds 

Fireworks AI intends to use the funding to expand its team and enhance its platform to enable developers to quickly move AI applications from prototype to production. 

Sonya Huang, General Partner at Sequoia Capital, said, "Fireworks AI is well positioned to lead this industry shift. Their team's expertise in building high-performance inference stacks and innovative approach to realizing composite AI systems will provide developers with scalable AI solutions that were previously only available to tech giants."

FireworksAI Doing what?

Led by CEO Joe Lin, Fireworks AI provides an inference engine to build production-ready composite AI systems. Using minimal manual data, ultra-fast LoRA fine-tuning enables developers to quickly customize models to their specific needs, transitioning from dataset preparation to querying a fine-tuned model in minutes. These models maintain the same performance and cost-effectiveness as the base model when deployed.

Since its inception, the company has been committed to providing developers with efficient reasoning about popular models. Today, the company offers more than 100 state-of-the-art models in text, image, audio, embedded, and multimodal formats, optimized for latency, throughput, and cost per token.

"Unlike proprietary large models that are generic, non-proprietary and difficult to customize, Fireworks AI delivers production-grade models that are smaller, privately and securely deployable," said Joe Lin, CEO. 

client base

AI startups like Cresta, Cursor, and Liner, as well as developers at digital native giants like DoorDash, Quora, and Upwork, use specialized models from Fireworks AI

Partnerships 

Since its debut, Fireworks has made significant upgrades to its platform and signed several agreements with providers in the AI stack. 

This includes working with Nvidia, AMD, Amazon Web Services, Google Cloud Platform and Oracle Cloud to optimize model infrastructure. The company has also partnered with MongoDB, a cross-platform NoSQL database, to create the Interactive Retrieval Augmentation Generation or RAG product, which helps bring authoritative, real-time data into large language models to improve accuracy and reduce errors.