Applied Compute raises $80M at $1.3B for custom-trained enterprise models

The round was led by Kleiner Perkins, with continued backing from Elad Gil, Lux, Greenoaks, Neo and Hanabi, and it lifts total funding to $160 million. The round was reportedly priced against about $12.8 million in annualized revenue. That implies a multiple above 100x, though the figure has not been confirmed by the company.

The startup was founded by three former OpenAI researchers, Yash Patil, Rhythm Garg and Linden Li. It came out of stealth in October 2025 with an $80 million raise from Benchmark, Sequoia and Lux. Each custom model is trained on the data of the customer that commissioned it. Early customers include DoorDash, Cognition and Mercor. In August, it was reportedly in talks to raise hundreds of millions at roughly a $3 billion valuation, led by Elad Gil. That round had not closed, and only The Information had the terms.

For founders, the valuation jump suggests investors now treat enterprise model customization as a separate infrastructure layer. It is no longer seen as a feature bundled with frontier models. Teams selling fine-tuning or agent tooling should expect direct competition from well-funded specialists. Applied Compute embeds engineers with client teams and delivers agents in days instead of months. That sets a speed benchmark for anyone pitching the same buyers.