Applied Compute, an artificial intelligence infrastructure startup, has raised $350 million in new funding at a $3.25 billion valuation. The round was led by investor Elad Gil, with participation from Kleiner Perkins, Lux Capital, NVIDIA, Shopify Ventures, HubSpot Ventures and other investors.
The company also announced that its annualized revenue run rate has surpassed $100 million, less than 18 months after its launch. Applied Compute provides infrastructure that helps companies train, deploy and improve AI models for their business needs.
Its platform supports large language model (LLM) training and inference, the process of running models to generate responses. Applied Compute also offers machine learning research support, helping customers with tasks such as evaluating newly developed models.
The company uses asynchronous reinforcement learning to allow data generation and model training components to run at the same time, helping speed up development. Its platform also supports self-distillation, a training method that removes the need for separate teacher models. Other features include an AI assistant for summarizing training results, an autoscaling engine for inference workloads and tools for routing requests to cost-efficient models.
Applied Compute’s platform includes AC2, its Agent Cloud for training, inference and continual improvement, as well as AC Inference and Shift-1, a model router for AI coding agents. Since July, customers have conducted more than 1,700 training runs on AC2. AC Inference now serves more than one trillion tokens per day, and 40% of customers use the platform for both training and inference.
The company said it has more than doubled the number of enterprise customers since its previous funding round. Customers include Microsoft, Shopify, DoorDash, Harvey, Mercor, Cognition, Clay and Decagon. Applied Compute has also reported results from customer projects, including lower inference costs for Harvey and Clay through custom model training and routing.
Founded by Rhythm Garg, Yash Patil and Linden Li, Applied Compute plans to use the new funding to expand its compute capacity, grow its team and accelerate enterprise deployments. The company says its broader goal is to build AI infrastructure that allows businesses to own and improve the AI systems powering their products and operations.
