bytevyte
bytevyte
Language
EN DE ES FR PT IT
Follow
ai-beats —

Decart Secures $300M at $4B Valuation to Scale Real-Time Decart World Models and Inference Stack

Decart world models

Decart has secured $300 million in a new funding round, valuing the AI research lab at $4 billion as it scales its real-time Decart world models. Radical Ventures led the investment, which included participation from NVIDIA, Adobe Ventures, Toyota Ventures, and eBay Ventures. The capital injection will support the development of the company's vertically integrated inference stack and its specialized model architectures designed for immediate, interactive AI experiences.

The core of the company's technical advantage is DOS 2.0, an infrastructure product that achieves over 1,600 tokens per second for agentic inference. This speed is critical for the Decart world models, which aim to simulate complex environments with zero latency. By controlling the full stack from the model layer down to the inference engine, the lab claims it can deliver performance levels that traditional, general-purpose cloud providers cannot currently match. This vertical integration allows for optimizations that are specifically tuned for the high-throughput requirements of video generation and physical simulation.

Strategic Impact of Decart World Models

The company is currently developing two distinct model lines to address different market needs. The first, Lucy, is optimized for real-time immersive gaming and interactive advertising. It generates high-fidelity video and interactive elements on the fly, allowing for digital environments that respond instantly to user input. This capability positions the lab as a key player in the next generation of digital media and entertainment infrastructure, where static assets are replaced by dynamic, AI-generated worlds.

The second model line, Oasis, focuses on physically accurate simulations. This technology is intended for robotics and autonomous vehicle development, where AI agents must understand and predict the physical consequences of their actions. By providing a high-speed environment for these agents to train in, the platform reduces the need for expensive and slow real-world testing. The involvement of Toyota Ventures suggests a strong strategic interest in these physical AI applications for future mobility solutions and industrial automation.

The investment from NVIDIA and Adobe further highlights the strategic importance of real-time inference. NVIDIA's participation indicates a focus on hardware-software co-optimization, while Adobe's interest likely stems from the potential for Lucy to transform creative workflows and generative media. As enterprises move from experimental chatbots to complex, real-time agents, the demand for specialized inference stacks like DOS 2.0 is expected to grow. The ability to process 1,600 tokens per second provides a significant lead over standard LLM providers who often struggle with the latency requirements of interactive video.

Decart plans to use the $300 million to expand its research team and increase its compute capacity. The company stated that its next milestone involves the public release of updated Oasis benchmarks, demonstrating improved physical reasoning in simulated environments later this year. This funding round places the lab among the most well-capitalized AI startups focusing on the intersection of generative video and physical world simulation, a sector that is becoming increasingly competitive as foundation model providers seek new ways to apply their technology beyond text and static images.

While we strive for accuracy, bytevyte can make mistakes. Users are advised to verify all information independently. We accept no liability for errors or omissions.

AI-generated image.

✔Human Verified

Share