AI cloud startup Applied Compute raises $350M after reaching key revenue milestone

Applied Compute Inc., a rapidly expanding artificial intelligence infrastructure provider, has secured $350 million in funding at a $3.25 billion valuation, the company announced on Monday. Led by prominent angel investor Elad Gil, the round saw participation from Nvidia Corp., Shopify Ventures, HubSpot Ventures, and over a dozen other investors.

Applied Compute operates a cloud platform for developers to train large language models and perform inference, with a team of machine learning researchers assisting customers with their projects. The software stack for LLM training usually consists of two main components, one for creating multiple model copies and assigning tasks, and another for refining weights using training data.

Applied Compute's platform employs asynchronous RL and self-distillation to enhance efficiency, eliminating the need for teacher models and automating repetitive tasks through an AI assistant. The platform also supports inference workloads with an autoscaling engine and online RL feature that fine-tunes LLM weights based on prompts.

Their model router, Shift-1, routes coding agent requests to the most cost-efficient LLM and can distribute complex tasks among multiple LLMs. With annualized recurring revenue surpassing $100 million within less than 18 months, Applied Compute, founded by Rhythm Garg, Yash Patil, and Linden Li, plans to expand its market presence and infrastructure using the raised capital. The company currently serves major tech firms such as Microsoft Corp. and DoorDash Inc.