Decart Raises $300M for AI Optimization Software

Decart.ai Inc. has successfully closed a substantial $300 million funding round, propelling its valuation to nearly $4 billion. The significant capital infusion was spearheaded by Radical Ventures, with crucial participation from industry giants Nvidia Corp. and Adobe Ventures, alongside Toyota Ventures and other institutional investors. Notably, the round also attracted prominent angel investors, including Andrej Karpathy, a co-founder of OpenAI.

This funding marks a pivotal moment for Decart, a company focused on accelerating AI development through its specialized software. The investment will fuel the enhancement of its product suite, particularly its flagship optimization platform, DOS, and its advanced world models, Lucy and Oasis. Decart's technology addresses a critical bottleneck in AI development: the time-consuming and costly process of optimizing neural networks for diverse hardware architectures. Traditionally, this can take months per chip, but Decart's DOS platform reportedly compresses this timeline to mere weeks, significantly reducing costs and enabling greater flexibility for AI workloads.

The company simultaneously unveiled DOS 2.0, an updated version of its core software. This iteration boasts impressive performance gains, enabling AI agents to process over 1,600 tokens per second, an eightfold increase over the current industry average. Furthermore, DOS 2.0 empowers world models to handle up to 100 frames of high-definition video per second, a critical advancement for applications requiring real-time visual processing. This enhanced capability is foundational to Decart's other offerings.

Decart's world models, Lucy and Oasis, leverage the power of DOS. Lucy is designed for real-time video stream manipulation, offering applications such as virtual try-on experiences for retail or rapid visualization of interior design concepts. A key innovation in Lucy is its ability to self-correct visual fidelity issues that can cascade across video frames, a common challenge in video AI. By training on its own imperfect outputs, Lucy can proactively identify and rectify errors, ensuring higher quality results. The platform also shows promise in generating synthetic training data for robotics, particularly for tasks like warehouse automation.

For more complex robotic applications, developers can integrate Lucy with Oasis, Decart's second world model. Oasis specializes in generating sophisticated 3D environments, such as simulated logistics warehouses, providing realistic digital twins for training and testing advanced robotic systems. The company plans to release new iterations of both Lucy and Oasis in the near future, building on the momentum from this significant funding round.

The substantial investment in Decart underscores the escalating demand for efficient AI development tools. As the AI sector continues its rapid expansion, projected to reach hundreds of billions of dollars in the coming years, companies that can streamline training, inference, and model deployment are poised for significant growth. Decart's focus on optimization and advanced world models positions it to capture a substantial share of this expanding market, particularly as enterprises grapple with the complexities of deploying AI at scale across diverse hardware and applications.