California-based AI startup Applied Compute is in talks to raise hundreds of millions of dollars in a funding round led by Elad Gil. The round would value the company at approximately $3 billion, more than doubling its $1.3 billion valuation from an $80 million round in April 2026. Founded by former OpenAI researchers, Applied Compute helps enterprise clients like DoorDash customize and run open-source AI models using their own data, reflecting a broader industry shift toward open-weight AI infrastructure.
Applied Compute $3 billion valuation
- ▪The proposed $3 billion valuation would more than double Applied Compute's previous valuation of $1.3 billion, which was established during an $80 million funding round disclosed in April 2026.
- ▪California-based startup Applied Compute is in negotiations to raise hundreds of millions of dollars in a funding round that would value the company at approximately $3 billion.
- ▪Applied Compute's previous $80 million funding round in April 2026 was led by Kleiner Perkins, bringing the company's total funding raised to approximately $160 million.
Open-weight AI infrastructure demand
- ▪Gartner forecasts that AI inference will account for 55% of worldwide AI-optimized infrastructure-as-a-service spending in 2026 as AI applications transition to production workloads.
- ▪The open-source AI platform Hugging Face reported reaching 13 million users and hosting over 2 million public models in 2025, signaling growth in the open-source AI ecosystem.
Agent Cloud platform capabilities
- ▪Applied Compute's Agent Cloud, or AC2, platform allows enterprises to train and deploy specialized AI models using their own data and workflows while keeping their agent infrastructure unchanged.
- ▪An Applied Compute experiment demonstrated that training a router to distribute software-engineering tasks among multiple AI models achieved GPT-5.5-level performance at a 25% lower cost.
Elad Gil led funding round
- ▪Applied Compute was founded by Stanford graduates Rhythm Garg, Linden Li, and Yash Patil, who previously worked on reinforcement learning and the o1 model at OpenAI.
- ▪Investor Elad Gil is reportedly leading Applied Compute's new funding round, having previously participated as an investor in the company's April 2026 funding round.
Enterprise model customization shift
- ▪Applied Compute focuses on running and customizing open-source model weights for enterprise clients, with DoorDash reported as a customer using the platform for custom agent work.
- ▪Applied Compute's previous $80 million funding round in April 2026 was reportedly priced against $12.8 million in annualized revenue.
Story comments
Loading comments…