Decart Raises $300M to Expand Real-Time AI Infrastructure and World Model Development

Decart, a San Francisco-based AI infrastructure company focused on real-time video and world models, has raised $300 million in funding. The funding round was led by Radical Ventures. New investors included NVIDIA, Atreides Management, Valor Equity Partners, Adobe Ventures, Toyota Ventures, and eBay Ventures. Existing investors such as Sequoia Capital, Benchmark, and Zeev Ventures also participated in the round. 

Individual backers included Andrej Karpathy, Michael Eisner, the Nintendo family, and gaming investor Moritz Baier-Lentz.

Decart Focuses on Real-Time AI Infrastructure

Founded by CEO and co-founder Dean Leitersdorf, Decart develops infrastructure designed for real-time and interactive AI applications.

The company’s proprietary Decart Optimization Stack (DOS 2.0) supports agentic inference speeds of more than 1,000 tokens per second. It also enables full-HD world model inference at up to 100 frames per second across different hardware systems, including NVIDIA GPUs, Google TPUs, and Amazon Trainium chips.

Decart positions its platform as infrastructure for production-scale AI applications that require low latency and high-performance processing.

AWS Partnership and Commercial Expansion

The company generates revenue through infrastructure agreements with cloud providers and AI labs.

Decart has also established a technical and commercial partnership with Amazon Web Services (AWS)  to run its Lucy2 model on Trainium3 hardware. According to the company, the deployment achieved more than 80% Model FLOPS Utilization.

Funding to Support Product Expansion

Decart plans to use the new funding to expand its three primary product lines.

These include Lucy, its world model platform for immersive experiences; Oasis, a world model platform designed for physical AI applications; and the Decart Optimization Stack.

The company also plans to accelerate product development and expand operations across New York City and Tel Aviv.