Applied Compute has raised $350 million at a $3.25 billion valuation in a round led by angel investor Elad Gil. The startup, which sells a cloud platform for training and running large language models, announced the deal after passing $100 million in annualized recurring revenue less than 18 months after launch.
Nvidia, Shopify Ventures and HubSpot Ventures joined the round, along with more than a dozen other investors. Applied Compute said it will use the money to grow its market presence and buy more infrastructure.
The valuation has climbed quickly this year. In April, Applied Compute disclosed an $80 million round at a $1.3 billion post-money valuation, bringing its total funding at the time to $160 million. By August, Gil was reported to be in talks to lead a new round at roughly $3 billion.
Key facts
- 01Applied Compute raised $350 million at a $3.25 billion valuation in a round led by Elad Gil.
- 02Nvidia, Shopify Ventures and HubSpot Ventures also invested, along with more than a dozen others.
- 03The company says annualized recurring revenue is above $100 million, less than 18 months after launch.
- 04Its April round raised $80 million at a $1.3 billion post-money valuation.
- 05Customers include Microsoft and DoorDash.
Revenue has moved at a similar pace. Applied Compute's annualized revenue was reported at about $50 million in August, so the $100 million mark means it has roughly doubled in a matter of months.
Applied Compute was founded in 2025 in San Francisco by Yash Patil, Rhythm Garg and Linden Li, all former OpenAI researchers. Patil is the company's CEO.
The platform is aimed at enterprises that want to customize and deploy open-source and open-weight models on their own data. It supports asynchronous reinforcement learning, which runs data generation and weight updates at the same time, and self-distillation, which removes the need for a separate teacher model. An online RL feature periodically fine-tunes a model on the prompts it receives.
On the serving side, Applied Compute offers an autoscaling inference engine and Shift-1, a model router that sends each request to the most cost-efficient LLM and can split complex tasks across several models. A built-in assistant summarizes what a model learned during training, and the company also staffs machine learning researchers who help customers build evaluations for new models.
Customers include Microsoft and DoorDash, which uses the platform for specialized agent workflows. Applied Compute's growth has been tied to enterprises seeking lower inference costs and less reliance on a single model vendor.
In a blog post announcing the round, the founders wrote that "Training, inference, and routing are the first pieces of a much larger stack," and that the company is building "the AI infrastructure every company will need to fully own their intelligence."
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
