Decart
Company Information
Headquarters
San Francisco, California, USA
Year Founded
2023
Industries
Artificial Intelligence, Generative AI, Media & Entertainment, Robotics, Computing Infrastructure, Simulation
What Is Decart?
Decart is an artificial intelligence company developing real-time AI infrastructure and world models designed to operate with very low latency. Its technology spans the computational stack from hardware-level optimization to generative models capable of transforming live video and creating interactive simulated environments. Decart’s core work centers on world models, generative video, and AI compute optimization, with applications ranging from digital media to physical AI systems such as robotics and autonomous vehicles.
What Does Decart Do?
Decart develops the infrastructure and AI models needed to make computationally intensive AI applications operate in real time. The company combines its Decart Optimization Stack (DOS) with specialized world models that generate or modify visual environments continuously rather than relying only on offline generation.
Its Lucy models transform live video as it is being streamed, while Oasis generates interactive environments for physical AI that respond to user or machine actions. Decart also works at the infrastructure layer, optimizing model training and inference across different AI accelerators. This combination supports applications including video editing, virtual try-on, advertising, gaming, robotics simulation, autonomous vehicles, and synthetic-data generation.
Decart Technology
Decart’s technology combines world models, real-time generative AI, and low-level compute optimization. Its research includes models that generate frames conditioned on actions, allowing an environment to change continuously in response to inputs. This approach is particularly relevant to interactive experiences and physical AI, where a model must respond quickly enough for an agent or user to interact with the generated environment.
Underneath these models is the Decart Optimization Stack (DOS), a vertically integrated training and inference platform. Decart says DOS works across hardware-aware model design, kernels, memory systems, proprietary compilers, and inference optimization. The company supports infrastructure including NVIDIA GPUs, AWS Trainium, and Google TPUs.
Decart also researches real-time video generation and editing. Lucy can manipulate individual elements of a video while it is playing, while Oasis generates responsive environments for simulation. Oasis 3 supports synchronized multiple-camera views and closed-loop interactions with physical control signals, providing real-time simulation for autonomous systems rather than only producing prerecorded synthetic video.
Why Decart Matters
Decart is relevant to an emerging shift from AI models that generate static outputs toward systems that can generate and respond to environments continuously in real time. That capability has implications beyond generative media because robots and autonomous systems need models capable of responding to actions and changing physical conditions with minimal delay.
The company is also addressing the computational requirements behind these systems through DOS rather than treating the AI model and infrastructure as separate layers. The combination of real-time world models and optimized AI infrastructure makes Decart relevant to two developing areas of AI: interactive generative experiences and physical AI simulation.
Decart Products & Platforms
Decart Optimization Stack (DOS)
DOS is Decart’s AI training and inference optimization platform and the infrastructure underlying the company’s models. It combines hardware-aware model design, kernel-level tooling, compiler technology, and inference optimization to improve how AI workloads use available compute. Decart has built DOS to operate across multiple accelerator platforms rather than being tied to a single hardware architecture.
Lucy
Lucy is Decart’s real-time video and world-editing model. It can modify people, objects, backgrounds, visual styles, and effects while video is being created or streamed. Applications include virtual try-on, live advertising, social platforms, streaming, and gaming. Decart reports that its current Lucy technology can perform continuous video transformations with very low latency.
Oasis
Oasis is Decart’s interactive world model for physical AI. Oasis 3 generates controllable simulated environments in real time so autonomous systems can interact with changing scenarios rather than train solely against predetermined simulations. It supports synchronized camera perspectives, physical control signals, and API access, with autonomous-vehicle simulation as its initial application.
Decart Applications
Decart’s technology spans both digital and physical applications. Lucy supports live video editing, virtual try-on, advertising, streaming, social media, and gaming, allowing visual content to change while it is being produced or viewed.
Oasis extends Decart’s world-model technology into robotics and autonomous-system simulation. Its generated environments can be used to expose autonomous systems to different geographies, weather conditions, traffic situations, and difficult edge cases without having to reproduce every scenario physically. Decart identifies autonomous vehicles as Oasis 3’s first focus, with broader potential applications across robotics, manufacturing, and drones.
Trending Companies
Frequently Asked Questions
What is the Decart Optimization Stack (DOS)?
DOS is Decart’s infrastructure platform for optimizing AI training and inference. It works across hardware-aware model architecture, kernels, memory, compilers, and inference systems and supports AI hardware including NVIDIA GPUs, AWS Trainium, and Google TPUs.
What is the difference between Decart's Lucy and Oasis models?
Lucy focuses on transforming and editing video in real time, while Oasis is an interactive world model designed to generate responsive simulated environments. Oasis is oriented toward physical AI applications such as autonomous vehicles and robotics, whereas Lucy targets video-based experiences including retail, advertising, streaming, social platforms, and gaming.
How can Oasis be used to train autonomous systems?
Oasis 3 lets developers define simulated environments and send physical control signals into them. The generated environment responds to those actions in real time, allowing autonomous systems to repeatedly experience different driving conditions, locations, weather, traffic situations, and rare scenarios without reproducing them in the physical world.
Can developers access Decart's models through APIs?
Yes. Decart provides API access to its models. Oasis 3 is API-accessible for interactive world generation, while Decart’s video APIs allow developers to stream video and receive AI-modified video in real time.
What can Decart's Lucy model change in live video?
Lucy can perform operations including character swaps, object addition or removal, background changes, style transformations, attribute changes, and visual effects while video is being streamed or generated.
Do You Have an Update for This Company?
Have an update, correction, or additional information about this company? Help us keep the futureTEKnow company database up to date by submitting the latest company information.
Related Companies
Latest Articles
OpenAI Astra for Law Brings GPT-6 Astra Into Legal Research and Law Firm Workflows
OpenAI’s Astra for Law combines GPT-6 Astra with specialized legal search, integrations, and governance controls. The launch shows how frontier
Are your AI projects stuck on your own data?
AI projects often stall because business data isn’t consistent or trustworthy. Here’s how to identify data gaps before they become
X Square Robot’s $276M Bet on Embodied AI and China’s General-Purpose Robots
X Square Robot has raised $276M from Xiaomi, Sequoia China, and other internet giants to scale its WALL-A embodied AI
EVAS Intelligence’s RISC-V AI Chips Land 1.5 Billion Yuan Series B to Power Global Expansion
EVAS Intelligence has raised 1.5 billion yuan to mass‑produce its RISC-V Epoch AI chips, deepen its full‑stack platform, and accelerate
Orkes Raises $60M to Orchestrate Enterprise AI Agents in Production
Orkes has raised 60 million dollars to turn its Netflix‑born workflow engine into a control plane for enterprise AI agents.
Sillage’s AI signal engine raises €1.7M to help enterprise sales teams follow the right buying signals
Paris-based Sillage has raised €1.7 million to launch an AI signal engine that helps enterprise sales teams follow the right
Cloneable’s agentic AI platform targets the infrastructure knowledge gap with $4.6M seed round
Cloneable is launching an agentic AI platform for infrastructure operations that captures institutional knowledge from retiring experts and turns it
Reliable Robotics Secures $160M to Scale Autonomous Aircraft Systems for Commercial Deployment
Reliable Robotics has secured $160M to scale production and deployment of its Reliable Autonomy System. This funding marks a pivotal
Ricursive Superintelligence’s $500M bet on self-improving AI raises the stakes in the AGI funding race
Excerpt: Ricursive Superintelligence has raised at least $500 million to build self‑improving AI, with GV and Nvidia backing a four‑month‑old
Bond AI raises US$2M to automate SME accounting in Brazil’s brutal tax maze
Brazilian startup BOND has raised US$2M to automate accounting for SMEs in Brazil’s complex tax system. Combining AI with human
Loop’s AI supply chain platform raises $95M Series C to predict disruptions before they hit
Loop just raised a $95M Series C to expand its AI-native supply chain platform, turning messy logistics data into early
London AI startup Lua wants to give every business its own AI workforce with $5.8M seed round
Linkedin X-twitter-square Facebook-square Startups & Business News AI agents are finally moving from demos to the day-to-day stack of real
futureTEKnow is focused on identifying and promoting creators, disruptors and innovators, and serving as a vital resource for those interested in the latest advancements in technology.
© 2026 All Rights Reserved.