The acquisition brings the Morpheus AI team to Roblox’s Foundation AI team, adding expertise in real-time video world models, a rapidly changing area of AI focused on generating interactive visual environments. Roblox says the deal supports its goal of delivering photorealistic multiplayer gaming.
This announcement follows Roblox’s April 2026 engineering update for Roblox Reality, which combines the Roblox Game Engine, Roblox Cloud, and video world model. Roblox also welcomed Joe Chen, founder of Dynamics Lab, and Alberto Hojel, founder of Lucid AI, who joined the company in late 2025.
Taken together, these moves indicate that Roblox is trying to solve three technical problems that still limit AI-generated interactive worlds: latency, consistency, and visual quality.
Morpheus AI founder Xun Huang wrote on LinkedIn that he joins Roblox after the Morpheus AI team worked on a video world model architecture that includes a self-enforcing and autoregressive video diffusion transformer known as AR-DiT.
Huang writes: “This work unlocks something unprecedented: the ability to go beyond offline pre-rendered AI video generation and instead simulate interactive worlds in real time.”
Morpheus AI team joins Roblox Reality effort
Roblox describes the acquisition of Morpheus AI as a strategic step toward its Roblox Reality vision. The company says the team’s efforts align directly with its plans to support photorealistic multiplayer gaming.
Anupam Singh, Roblox’s senior vice president of engineering, said in a separate engineering post that the path to interactive AI must address “latency, consistency, and quality.”
Roblox says Huang’s Self Forcing effort aims to move video models away from slow offline generation and toward a faster, autoregressive, interactive generation engine. Morpheus AI was founded in August 2025 to pursue real-time responsiveness for video world models.
Huang wrote on LinkedIn that Morpheus AI is working to bridge the “gap between deterministic game engines and generative world models.”
He added, “Joining Roblox means our technology will help power the experiences delivered to millions of players every day.”
According to Roblox, Morpheus AI’s models are expected to support the evolution of the Roblox video model, also known as the Super Upsampler. In the Roblox Reality architecture, the video model handles visual components such as high-fidelity textures, realistic lighting, and fluid physics, while the Roblox engine manages multiplayer state, game logic, and synchronization.
Roblox adds founders of Dynamics Lab and Lucid AI
The deal with Morpheus AI is part of Roblox’s broader AI adoption and technology push.
Joe Chen, founder of Dynamics Lab, joined Roblox in late 2025. At Dynamics Lab, Chen and his team worked on a real-time generative world engine that allows users to upload images, such as photos, concept art, paintings, and drawings, and turn them into interactive worlds that can be modified through text prompts.
Lucid AI founder Alberto Hojel also joined Roblox in late 2025. According to Roblox, Hojel’s work focused on making AI-generated worlds more believable by combining video world models with deterministic game logic.
According to Singh, Hojer invented the Roblox game cartridge harness, which extends the video world model with the logic of a proprietary game engine. The aim is to transform the AI-generated world from a visual environment into an interactive multiplayer experience with consistent progression and rules.
Roblox assembled this combined effort as part of a hybrid approach. Standalone video world models can generate visually rich environments, but currently lack long-term memory, consistent logic, user input controls, and reliable multiplayer states, Roblox said.
Hybrid architecture combining engine logic and AI video
Roblox Reality is designed to split responsibility between the Roblox game engine and the video world model.
The Roblox engine manages structured simulations such as state synchronization, collisions, rules, player position, and persistent world data. Roblox video models add generative visual details such as lighting, textures, secondary motion, and natural environment effects.
Roblox says the goal is to support high-fidelity multiplayer worlds without forcing creators to manage the cost and complexity typically associated with photorealistic graphics. The company’s latest engineering post states that Roblox Reality’s goal is to reach 4K resolution at 60Hz.
To support that effort, Roblox says it will move Roblox video models to adjacent edge data centers equipped with H200/B200-class GPUs or equivalent accelerators.
Sunil Rao, managing director of Roblox India, wrote on LinkedIn that Roblox “fuses generative video world models directly into structured game engines.”
He added: “We are moving from rendering games to simulating reality in real time.”
The initial version of Roblox Reality is expected to be released later this year or early next year. Roblox says it still faces computing and scaling challenges before the technology can support millions of simultaneous players, but the acquisition of Morpheus AI and the hiring of its founder will give Roblox more specialized talent for the next phase of its photorealistic multiplayer creations.
