Fireworks AI Raises $52M in Series B to Boost GenAI Inference Platform

Fireworks AI is revolutionizing the way AI applications are developed and scaled by providing a powerful inference engine optimized for various AI model formats, including text, image, audio, embedding, and multimodal.

Fireworks AI Raises $52M in Series B to Boost GenAI Inference Platform
Source: Fireworks AI

Company Name: Fireworks AI
Location: Redwood City, CA, USA
Sector: Technology, Artificial Intelligence
Funding Details: Raised $52M in Series B funding, valuing the company at $552M. The round was led by Sequoia Capital, with participation from NVIDIA, AMD, and MongoDB. The total capital raised by the company now stands at $77M. Previous investors include Benchmark, Databricks Ventures, and high-profile individuals such as former Snowflake CEO Frank Slootman, former Meta COO Sheryl Sandberg, and others.

Purpose of Investment: To expand the team and enhance the platform, enabling developers to efficiently transition AI applications from prototype to production.

Leadership: Led by CEO Lin Qiao.

Product: Fireworks AI offers an inference engine that enables the building of production-ready, compound AI systems. The platform features ultra-fast LoRA fine-tuning that minimizes the need for human-curated data, allowing developers to quickly customize AI models to specific needs—from dataset preparation to querying a fine-tuned model within minutes. The deployed models maintain performance and cost efficiency, comparable to their base models.

About Company: Fireworks AI is revolutionizing the way AI applications are developed and scaled by providing a powerful inference engine optimized for various AI model formats, including text, image, audio, embedding, and multimodal. The company supports over 100 state-of-the-art models, optimizing them for latency, throughput, and cost per token. Its solutions are used by developers at AI startups and digital-native giants like DoorDash, Quora, and Upwork, who rely on Fireworks AI for specialized, efficient, and scalable model deployments.