HANGZHOU, China, May 26, 2026 /PRNewswire/ — Manycore Tech today announced the open-source release of Aholo Viewer, a 3D Gaussian viewer that allows users to seamlessly explore massive 3D environments directly from a web browser. Fully compatible with smartphones, PCs, and VR devices, the viewer requires no client installation and can smoothly render city-scale 3D scenes with over 1 billion Gaussian splats — an area roughly equivalent to the entirety of Hangzhou’s West Lake.
Building the Infrastructure for Mass Adoption of the 3D Internet
“For more than 30 years, the internet has been largely confined to 2D formats such as text, images, and video,” said Huang Xiaohuang, co-founder and chairman of Manycore Tech. “A room is reduced to a handful of photos, a product becomes a flat image. Rich physical-world information has continuously been compressed into lower-dimensional representations. By open-sourcing core 3D infrastructure like Aholo Viewer, we aim to accelerate the arrival of the 3D internet era.”
Aholo Viewer employs a streamable Level-of-Detail (LoD) architecture, enabling 3D content to be transmitted across devices as smoothly as video streaming. Users can instantly access and interact with billion-splat 3D scenes directly from a browser, regardless of hardware platform.
Industry observers note that the number of Gaussian splats a browser can handle, along with rendering speed, is a key indicator of its practicality. Fei-Fei Li’s open-source Spark 2.0 engine can stream over 100 million Gaussian splats on consumer-grade devices. Aholo Viewer, by contrast, supports up to 1 billion splats — a tenfold increase — while delivering meaningful improvements in rendering speed, fluidity, and memory efficiency.
High-performance 3D browsers are expected to become the primary gateway for mainstream users to consume 3D content. In practical terms, exploring immersive 3D environments may eventually feel as effortless as scrolling through short-form videos today — a single link, and users can “step into” any real or virtual world.
From “Viewable” to “Usable”: Expanding the Role of Spatial Intelligence
For Manycore Tech, enabling 3D content to be viewed is only the first step. Equally important is making it usable. Alongside the open-source release of Aholo Viewer, the company is also opening APIs for its Aholo Spatial Intelligence Platform — including spatial reconstruction, cloud rendering, and AI-powered 3D model generation — to tackle the bottlenecks in producing, circulating, and deploying 3D content.
With these capabilities, 3D scenes can evolve from passive viewing experiences into editable, programmable digital assets that integrate directly into business workflows. Tourism operators could rapidly reconstruct entire scenic destinations, filmmakers could generate virtual production environments, and game developers could design interactive experiences inside faithfully reconstructed real-world spaces.
By combining open-source visualization infrastructure with open APIs for spatial computing, Manycore Tech is systematically building the foundational infrastructure for the emerging 3D internet ecosystem — staying true to its role as a core infrastructure provider for the spatial intelligence industry.
The Long-Term Value of the 3D Internet: Fueling AI’s Understanding of the Physical World
The significance of the 3D internet extends far beyond the evolution of online content from 2D to 3D. Its deeper importance lies in addressing one of artificial intelligence’s most fundamental limitations: the inability to form a complete spatial understanding of the physical world.
Historically, the internet has been dominated by 1D and 2D data, forcing AI systems to learn primarily from flat representations. This helps explain why today’s AI models can generate sophisticated written content yet still struggle with spatial reasoning, distorted visual outputs, and basic real-world physical tasks.
As core 3D internet infrastructure such as Manycore Tech’s Aholo Viewer becomes more widely adopted, long-standing barriers to the distribution and accessibility of 3D content could be substantially reduced. If creating and consuming 3D content becomes as effortless as sharing short videos today, vast amounts of real-world spatial data may emerge at scale — providing critical training data for embodied AI systems, including robotics and other intelligent agents designed to operate in physical environments.
