Elon Musk's xAI Ventures into 'World Models' for Gaming and Beyond

Elon Musk's xAI joins race to build 'world models' to power video games

Elon Musk's xAI is pushing to build so-called world models, joining rivals such as Meta and Google in the race to develop artificial intelligence systems that can navigate and design physical environments. The San Francisco-based start-up hired specialists from Nvidia over the summer to work on these next-generation AI models, which train on videos and data from robots to understand the real world. World models could push the capabilities of AI beyond that of the large language models, trained on text, that underpin popular AI tools such as ChatGPT and xAI's Grok. Two people familiar with the plans said the company was building world models with a view to applying them in gaming, where they could be used to generate interactive 3D environments. One of the people added that they could be applied to AI systems for robots. xAI has hired Zeeshan Patel and Ethan He, two AI researchers from Nvidia with experience in world models. Nvidia has been a leader in developing this technology with its Omniverse platform, which creates and runs simulations. Some tech groups have vast expectations of world models, which could unlock uses for AI beyond software and computers in physical products such as humanoid robots. Last month, Nvidia told the Financial Times that the potential market for world models could be almost the size of the present global economy. xAI would release a "great AI-generated game before the end of next year", Musk said in a post on X, confirming a target the billionaire set last year. On Tuesday, xAI launched its latest image and video generation model, which it said had "massive upgrades" and is free to use. Current video generation models, such as OpenAI's Sora, generate frames of images for videos by predicting patterns learned from training data. World models would be a big advance as they would have a causal understanding of physics and how objects interact in different environments in real time. The company is advertising for technical staff in both image and video generation to join its "omni team", which "creates magical AI experiences beyond text, enabling understanding and generation of content across various modalities, including image, video and audio". Salaries for these jobs range from $180,000 to $440,000. It also has an open position for a "video games tutor", who will train Grok to produce video games and enable "users to explore AI-assisted game design", for $45 to $100 an hour. Musk follows other leading AI labs, such as Google and Meta, that are also working on these systems. However, world models remain a huge technical challenge. Finding sufficient data to simulate the real world and to train such models has proved difficult and costly. Michael Douse, head of publishing at Larian Studios, which develops the video game Baldur's Gate 3, said on X this week that AI could not solve the "big problem" for the games industry, which is "leadership [and] vision". He added that the industry did not need "more mathematically produced, psychologically trained gameplay loops [but] rather more expressions of worlds that folks are engaged with, or want to engage with". xAI, Patel and He did not respond to requests for comment.

Elon Musk's xAI eyes gaming and robots in push toward real-world AI understanding

Serving tech enthusiasts for over 25 years. TechSpot means tech analysis and advice you can trust. Looking ahead: Elon Musk's artificial intelligence startup, xAI, is ramping up its efforts to develop a new generation of AI systems known as world models - technologies designed to interpret and construct representations of the physical world. The initiative places xAI alongside major competitors such as Google DeepMind and Meta in the emerging race to build AI that can reason about real-world environments, not just text or images. People familiar with the company's plans told the Financial Times that xAI recently hired Nvidia researchers Zeeshan Patel and Ethan He, both of whom have experience developing world models and simulation-based AI. Nvidia, one of the largest suppliers of computing hardware for AI research, has invested heavily in its Omniverse platform, a suite of tools for building and operating industrial-scale simulations. Some of that expertise is now shifting, in part, to Musk's venture. World models represent one of the most ambitious frontiers in artificial intelligence research. Unlike large language models that generate text based on statistical patterns in data, world models are designed to learn how physical systems behave. They train on vast collections of video, sensor, and robotic data to build a causal understanding of objects, motion, and environmental dynamics. By modeling the physics of real or simulated environments, these systems could enable AI to act more intelligently and safely in the physical world, an essential step toward autonomous machines and advanced robotics. xAI's push into world models is expected to extend beyond research and into commercial applications. Two people familiar with the project said the company sees potential for the technology in video game development, using it to automatically generate interactive 3D environments. One of the sources added that xAI also views the models as essential for future robotic systems capable of navigating complex spaces. Such capabilities could move AI beyond purely software-based tasks, establishing a foundation for physically embodied intelligence. Musk has outlined some of xAI's near-term goals, including plans to release what he described as a "great AI-generated game" before the end of next year. The project would align with the company's broader effort to expand its generative tools beyond text and images. Last week, xAI unveiled its latest image and video generation model, describing it as a major technical upgrade and making it freely available to the public. While video creation platforms such as OpenAI's Sora produce image sequences through statistical prediction, xAI and other researchers are attempting to train AI systems that understand cause and effect, or how one action leads to another in real time. The company has also begun advertising new technical roles in image and video generation for what it calls its "omni team," which describes its mission as creating "AI experiences beyond text" across visual and audio formats. Salaries for these positions range from $180,000 to $440,000, according to job postings. xAI is additionally recruiting a "video games tutor" to train its chatbot, Grok, in game design, offering hourly rates between $45 and $100. The broader promise of world models has drawn increasing attention across the AI industry. Nvidia told the Financial Times last month that the eventual market for these systems could approach the scale of the global economy, given their potential applications in manufacturing, robotics, and simulation-based training. Still, developing them remains a daunting technical challenge. Building accurate representations of the real world requires vast amounts of multimodal data and immense computing power, both of which are highly resource-intensive hurdles. Skepticism persists among segments of the gaming community, where world models are seen as unlikely to replace human creativity. Michael Douse, head of publishing at Baldur's Gate 3 developer Larian Studios, wrote on X that AI tools will not fix what he called the industry's "big problem," pointing instead to a lack of "leadership [and] vision." He argued that the future of games lies not in "mathematically produced, psychologically trained gameplay loops" but in artistic engagement and world-building.

Elon Musk wants AI to build games and poaches from Nvidia, but I'm skeptical

Merging AI with game engines isn't just a gamble on code but the future of creativity itself. What's happened? Elon Musk's xAI has intensified its push into gaming AI by recruiting ex-Nvidia specialists to build 'world models'. These world models are AI systems that can simulate real-world physics and environments. It will create game worlds that are AI-generated, physics-aware, and possibly reactive to player actions. Two Nvidia veterans, Zeeshan Patel and Ethan He, have been hired by xAI for world modeling. Musk has already announced that xAI will deliver a 'great AI-generated game' by the end of 2026. This is important because: It's not another "Elon does something bold" story. What's happening here strikes at the heart of how games are made and what it means when AI tries to replace parts of that. If world models can generate terrain, physics, and NPCs automatically, the role of designers and storytellers might shrink. Game devs could feel pressure too, especially in studios chasing cost savings by outsourcing to AI. AI slop: AI can generate a lot of game content, but will it be engaging, coherent, and meaningful? As a gamer, I am concerned that it could be shallow, repetitive, or glitchy. If parts of a game are generated, who gets credit? The control shifts toward those who own the models, infrastructure, and data, and not the human creators. With more AI-generated games, will there still be space for indie developers? Recommended Videos Why should I care? You may not be a game developer, but if you play games or simply follow tech, this matters more than you think. Letting AI build games might sound futuristic, but video games thrive on human direction, storytelling, and emotional design. Current generative tools still struggle with consistency and that spark of intentional creativity, so I'll keep my skepticism handy until we see something that feels less like a demo and more like a genuine gaming experience.

Elon Musk is 'full of crap' for claiming xAI will make an AI-generated game in a year, says Glen Schofield: 'I want to actually say that to him'

The Dead Space director embraces AI use in game development, but hasn't drunk the xAI kool-aid. Dead Space creator and former Striking Distance boss Glen Schofield is an advocate for AI use in game development and is not shy about it. During a keynote speech at Gamescom Asia x Thailand Games Show titled 'Ten Ways I Come Up with Ideas', Schofield championed generative AI as a brainstorming tool in game development, especially Midjourney. During the discussion he described how he uses Midjourney to experiment with visual ideas in the pre-concept phase, and also how he uses generative AI chat to brainstorm. Schofield -- who was speaking during the business-centric part of the convention -- says AI's growth in gamedev is basically inevitable. "History has shown us that every major technological leap creates whole new industries, opportunities, and specialists," he said. "AI is a technological leap -- it's like the PC, internet or cellphone, and just like those it's going to create another wave. It'll change the way we work and the way we do things. Yes, some jobs are going to shift, they may go away, but new ones -- millions of them -- will be created over the years. "Learn some AI that makes you better, faster and more efficient," he continued. "That's how you're going to make yourself indispensable. Now is the inflection point." Of course, without a time machine it's impossible to say whether Schofield is right, but it increasingly seems obvious that AI use will continue ramping up. But he's relatively pragmatic, clear-eyed even, compared to others. In an interview after his keynote chat, I asked Schofield what he thought of Elon Musk's claim that xAI will release a completely AI-generated game by the end of next year. He hadn't heard about Musk's claim but was unequivocal. "A year?" He said, visibly shocked. "In one year they'll be able to make a game? No, he's full of crap." He continued: "Here's some of the problems I see: you have AI companies making these AI tools and saying 'this is for world building' and 'this is for videogames', and I'm like: can you tell me, did you have any [game] developers help you? Did you have input from game developers? Because one thing game makers hate is when somebody outside the industry makes a tool and then we get it and it's so screwed up because they don't know how to make games. That has to happen first. I know a couple of companies that are [making AI-centric gamedev tools], but that has to happen first. "Making a game in a year? I mean look, somebody's going to do it, right. But I don't think it's going to be great yet. He's full of crap. That's what I believe. I want to actually say that to him." While Schofield's AI discussion was focused on the technology as a brainstorming tool, I asked him what he thought it might be good for in five years, or further afield, when it came to active game development. He mentioned world buildings and character building, "maybe taking it from a concept and building it out", but conceded that both "are still problematic" on the level of execution, citing the difficulty of nailing 3D character movement. Schofield spoke at length about using Midjourney to create early visual ideas that he would then hand over to an art director. He described taking old drawings of his own and splicing them together towards strange results, or just prompting bizarre specimens ("give this monster spaghetti hair, that's spiked"). Some 80 pieces of his (non-AI) art were exhibited at a gallery in San Francisco last year. While AI image generation draws from countless millions of images already on the internet, Schofield doesn't think it should be an ethical obstacle. "This is just my own thing: I don't use other people's art. Now, artists are gonna go 'alright [generative AI image creation] is stealing art' -- but it's not stealing art. I said to a couple of them: OK you have AI right here and you don't want to use other people's art, I get that. But: you photo bashed for the last 10 years. You're using other people's work: How is that any different? "'Well, I'm using my own style' [they might say], or something like that, and I think that's BS too. Because when you're an artist learning to draw, you're copying other people's stuff all the time, you're taking pieces. So what I would like to get out there: put this legal stuff behind you. Some of it's ethical, and everybody's got different ethics, so work with your own ethics, but do not ignore AI."

Elon Musk has been snapping up ex-Nvidia specialists to help make AI world models, which is fine but please keep away from our games

If there's one kind of AI that's most relevant to gaming, it would be the so-called 'world model.' This is a system that can learn to simulate real-world things, which might allow for 'easy' creation of realistic games. And that seems to have been on Elon Musk's mind, as his xAI has been hiring Nvidia specialists to work on just such AI models. That's according to the Financial Times, which cites "two people familiar with the plans" and says the two new hires, Zeeshan Patel and Ethan He, have experience working on world models. Apparently, the creations are intended to help create game environments. That tracks, too, given Musk's previous comments: "The XAI game studio will release a great AI-generated game before the end of next year." If these world models are put towards game world creation, I'm sure we'll have a whole bunch of new AI slop to look forward to. I mean, just look at the aforementioned X post -- Musk is responding to a clip of a game that looks absolutely atrocious. I'm sure it will allow for lots of bespoke and custom game scenes, of course. We've seen hints of the capabilities on this front with Google DeepMind's Genie 3, which allows you to interact with real-world environments of your choice, concocting some pretty impossible scenes. This sounds fun, but it can hardly compare to a well-crafted -- humanly-crafted -- game with vision and purpose. At least, in this writer's humble opinion. Though it's worth noting, xAI's world models probably won't be only for gaming. The Financial Times says that one of the sources claims they could be used in AI systems for robots. This would make sense, especially in the context of autonomous vehicles. Nvidia's own world model project, Nvidia Cosmos -- and remember, the new xAI world model hires were from Nvidia -- is used "to accelerate physical AI development for autonomous vehicles (AVs), robots, and video analytics AI agents." Musk's Robotaxi might be an obvious use for a world model: It would give AV software a safe way to learn without actually getting out there in the real world. Let's hope it's just that, eh? I don't know if I'll be able to stomach fully AI-made games -- certainly not the initial ones, which are bound to be the worst ones.

Dead Space creator and ex Call of Duty boss Glen Schofield says AI is "the right investment" and downplays theft concerns: "Picasso said 'good artists always borrow, great artists steal'"

"Use it. Let the lawyers figure it out. It's going to be figured out someday." Former Call of Duty lead and Dead Space creator Glen Schofield is going all in on AI. Speaking to VGC at GamesCom Asia, Schofield gave some of his ideas to "save" the games industry. "First, let's train everybody [with AI]. We know this new wave is coming." He adds, "people will say 'well, the software isn't ready yet'. If you're telling me about it, it's ready. I mean, ready to play with at least. You play with it. You learn it. You grow with it. Right?" He urges companies, "if everybody would just train their people... It's not a huge investment, it's the right investment." When it comes to concerns about Generative AI stealing from actual artists and creators, Schofield notes, "Every artist, every concept artist, is already doing photobashing. They've been doing it for 5-10 years. That's not your work." He adds, "that same person... was the one who said 'I don't want them taking work', and all that, and I'm like, you gotta be kidding me." Schofield, who has created artwork for games himself, admits he was concerned about this at first. However, his mindset is currently that "Picasso said 'good artists always borrow, great artists steal'." For developers who are not keen to steal others' work, Schofield suggests "make your own rules on how you want to deal with it," citing that he uses his own paintings and drawings to create a model for his prompts. But even without doing all that, he says, "use it. Let the lawyers figure it out. It's going to be figured out someday." This is antithetical to Hideo Kojima, who recently commented on his own feelings on AI, saying that he thinks "of AI as more of a friend," with it helping handle the tedium of development to "lower cost and cut down on time" as opposed to having it create ideas and assets (again, stolen from others' work). Clearly, AI continues to divide and will do so for a while yet.

Elon Musk's xAI hires Nvidia experts to develop advanced 'world models' - The Economic Times

Elon Musk's AI startup, xAI, is developing advanced "world models" that help AI understand and interact with physical environments. To do this, xAI hired experts from Nvidia experienced in training AI with videos and robot data. Nvidia leads this field with its Omniverse simulation platform.Elon Musk's AI startup, xAI, is working on developing "world models", according to the Financial Times. These are advanced AI systems that can navigate and design physical environments. These models aim to take AI beyond large language models like Grok and ChatGPT. To compete with companies such as Meta and Google, xAI hired experts from Nvidia over the summer. These specialists bring experience in training AI using videos and robotic data to help machines better comprehend the real world. As per the report, the experts recruited are Zeeshan Patel and Ethan He, both AI researchers from Nvidia who have worked on world models. Nvidia itself has been a leader in this field through its Omniverse platform, which builds and runs simulations. The report said that many in the tech industry have high hopes for world models, as they could enable AI to move beyond software and into physical products like humanoid robots. The company is also hiring technical staff for its "omni team," which focuses on creating AI systems that can understand and generate content across various formats, including photos, videos, and audio. Musk recently reiterated a goal he set last year, saying xAI will release a "great AI-generated game before the end of next year." Last month, xAI launched aits new image and video generation model, Grok Imagine, on social media platform X and offering it free to users. While current video generation models, like OpenAI's Sora, produce video frames by predicting patterns from training data, world models represent a significant leap forward. They would have a real-time, causal understanding of physics and how objects interact in different environments.

Elon Musk's xAI snaps up Nvidia talent to power next-gen AI gaming

The company recently recruited Zeeshan Patel and Ethan He, two AI researchers who previously worked on Nvidia's Omniverse platform. We just got the news that Elon Musk's AI startup xAI is quietly assembling a team of ex-Nvidia specialists to accelerate the creation of "world models," that is, advanced systems capable of understanding and designing virtual environments. The hires, revealed by the Financial Times, include Nvidia researchers Zeeshan Patel and Ethan He, both veterans in AI simulation, signaling real progress toward Musk's next big leap: a fully AI-generated video game, built without human intervention, which he claims will debut before the end of 2026. * You might be interested: Elon Musk promises fully AI-generated game by next year. World models go beyond language-based systems like ChatGPT or Grok, using visual and robotic data to learn how objects behave in real time, the same technology that could one day power autonomous vehicles and robots. And while the gaming industry remains skeptical, Musk's aggressive hiring spree (and xAI's recent upgrades in image and video generation) show he's serious about pushing AI into uncharted territory. What do you think about Musk's latest venture? Good or bad?

Elon Musk's xAI is Creating AI-Generated Video Game Built on Nvidia's World Model Tech

Inside xAI's AI-Generated Game: Nvidia Tech Brings Real-World Physics to Digital Spaces xAI, an artificial intelligence company led by Elon Musk, is developing an AI-generated video game that will be powered by world model technology. According to a recent Financial Times report, the firm has recruited former Nvidia researchers to accelerate the creation of next-generation AI systems capable of understanding and interacting with real-world physics. These systems differ from current large language models, which are primarily trained on text and images. World models are concerned with interpreting movement, cause and effect, and object behavior within three-dimensional worlds. The technology will grant AI a physical reasoning ability, allowing it to mimic real-world interactions and change the approach to constructing and experiencing digital space. Sources familiar with indicated that the company's early use cases for world models could include gaming, where AI would generate adaptive and realistic 3D environments. Over time, the same systems could support robotics applications, enabling machines to comprehend spatial dynamics better and design real-world spaces.

xAI is building AI models for gaming and robotics: Here's what we know

Next-gen AI models could revolutionize video games and autonomous robots Elon Musk's artificial intelligence startup, xAI, is moving beyond large language models and into a new frontier of machine intelligence - one that doesn't just talk about the world but understands it. The company is now building "world models", a next-generation class of AI systems that can learn the physics of reality, simulate dynamic environments, and even design interactive 3D spaces for gaming and robotics. Also read: OpenAI vs. Elon Musk: Lawsuit escalates with claims of intimidation against AI safety advocates Unlike the text-based large language models that underpin systems such as ChatGPT or xAI's own Grok, world models are trained on videos, physical simulations, and data from robots. Their goal is to build a causal understanding of how the world works, how objects move, how light behaves, how a hand might grasp a cup or a car might skid on ice. In short, these models aim to give AI a form of spatial and physical common sense. They could allow digital agents to navigate real or virtual environments with the same kind of intuitive logic that humans take for granted. Musk has said that xAI plans to release a "great AI-generated game" by the end of 2026, a statement that, like many of his pronouncements, straddles the line between ambition and provocation. According to reporting by the Financial Times, xAI has hired several former Nvidia specialists, including Zeeshan Patel and Ethan He, to work on this initiative. Nvidia, whose Omniverse platform already powers high-fidelity 3D simulations, has described the market potential for world models as "almost as large as the global economy itself." Also read: Inside Intel's reboot: Sachin Katti's blueprint for an open, heterogeneous AI future For xAI, video games are not just entertainment, they're testbeds for world models. Games provide complex but controlled virtual environments where AI can safely learn cause and effect, decision-making, and spatial reasoning. If successful, these models could later be extended to robotics, enabling autonomous systems that interact intelligently with the physical world. Earlier this week, xAI launched a new image and video generation model, touting "massive upgrades" and free access. But the company's job postings suggest an even broader vision. Its so-called "omni team" is hiring researchers across image, video, and audio generation to create "magical AI experiences beyond text." Salaries for these roles reportedly range from $180,000 to $440,000, with one listing for a "video games tutor" - a human who will train Grok to produce video games and help users explore AI-assisted design. That last detail hints at a future where game development could become an interactive collaboration between human designers and generative AI systems capable of shaping worlds in real time. The concept of a world model isn't new, researchers like David Ha and Jürgen Schmidhuber have been exploring it for years. But the surge in compute power, coupled with breakthroughs in multi-modal learning, has made it newly viable at scale. World models go beyond frame-by-frame video prediction (as seen in OpenAI's Sora) by understanding rather than merely imitating motion. Instead of stitching images together, they model the laws of physics, allowing for consistent and reactive environments that can be navigated, altered, and reasoned about dynamically. That makes them valuable not only for games and robots, but also for digital twins, simulation-based training, and next-generation AI assistants that can reason about the physical world, from traffic patterns to manufacturing processes. Not all voices in the gaming industry are enthusiastic. Michael Douse, head of publishing at Baldur's Gate 3 studio Larian, wrote this week that AI won't fix what he calls the industry's "big problem" - a lack of leadership and creative vision. "We don't need more mathematically produced gameplay loops," he said, "but more worlds that people actually want to engage with." That skepticism underscores a deeper question: will AI-generated worlds feel as rich and meaningful as those built by human hands? For now, xAI's ambitions are clear - to give AI the power to understand and shape the worlds we live and play in. Whether those worlds will feel alive is another challenge entirely.