ByteDance, the parent company of TikTok, will unveil an artificial intelligence model that generates three-dimensional virtual spaces in real time as early as next month. The move marks a full-scale entry into the "world model" business dominated by Meta and Alphabet, and is expected to intensify the battle for supremacy across the robotics, autonomous driving and extended reality ecosystems.
ByteDance founder and Chairman Zhang Yiming is personally directing the development of the next-generation spatial video AI model, mobilizing the company's internal AI capabilities and computing resources, according to Bloomberg and other foreign media on the 7th. The model, expected to be announced as early as next month, was built on Seedance, ByteDance's existing video generation AI. When a user issues a command by voice or movement, it creates an interactive 3D virtual world in real time in just 0.05 seconds. The model is expected to be released to the public through a partnership with Pico, ByteDance's virtual reality device manufacturing subsidiary.
Known as a world model, the technology allows AI to learn how the physical world works and then generate on its own a three-dimensional virtual environment that responds to a user's actions in real time. It can be used for short-form dramas, live streaming and 3D game video production. It can also be applied to simulation training for self-driving cars, robotics and AI agents. Google earlier showcased a world model under the name Genie.
ByteDance has chosen to handle 3D spatial computing in the cloud rather than on the user's device. Apple's Vision Pro and Meta's Quest series have struggled to reach the mass market because of the high price of their high-performance hardware and a shortage of content. Processing computation in the cloud, as ByteDance does, can sharply lower the specifications required of personal devices.
Bloomberg said that if ByteDance's new model succeeds, it could open a new front in competition with Meta and Apple, both of which have invested heavily in VR devices. Meta has worked to build a mass-market ecosystem around its Quest headsets, while Apple has marketed the Vision Pro as a premium spatial computing platform, but both companies have struggled to achieve mainstream adoption.
ByteDance, meanwhile, recently secured a large loan of $30 billion (about 40 trillion won) and is betting the company on investment in AI data center infrastructure and cutting-edge semiconductors.
