By Brent Li
As AI continues to reshape digital content creation, the battleground has expanded from text and 2D images into the three-dimensional space. VAST, an AI startup founded in 2023, has completed a $50 million Series A funding round co-led by Alibaba and SAIC Motor’s Hengxu Capital with participation from Baidu Ventures and other institutions.
This is the first time Alibaba and Baidu have simultaneously placed bets in the AI 3D niche, reflecting the consensus on 3D generation among top internet and industrial capital. VAST — which focuses on general 3D large models and interactive content platforms — plans to use the funds to expand its talent pool, iterate its core algorithms, and accelerate the development of user-generated content platforms.
Transforming hours into seconds
At the core of VAST’s offering is Tripo Studio, a platform that claims to have gathered over 6.5 million creators and generated nearly 100 million 3D models. The company recently launched new model architectures, including the Tripo P1.0, which can generate a functional 3D asset in under two seconds.
To embed itself into professional ecosystems, VAST has integrated Tripo via plugins for standard industry software including complex modeling tools such as Blender and Maya as well as videogame engines like Unity and Unreal. Tripo has also integrated with Comfy UI, a popular graphical interface for managing complex AI generation pipelines.
From miniature trains to coral reefs
The commercial applications of AI 3D generation are moving from the virtual to the physical world. In the UK, miniature model company Modelu is using Tripo to augment its traditional 3D scanning workflows. By using text prompts based on historical photos, the team generates varied heads and clothing for railway figurines, allowing the company to rapidly expand its product catalog without needing to re-scan physical actors.
AI 3D generation is also being deployed for environmental conservation. In Pujada Bay, Philippines, researchers are using 3D printed clay reefs to restore coral reefs. A team used Tripo to create complex, modular 3D structures based on a natural “gyroid” shape. The resulting high-surface-area geometries, ideal for coral attachment, were then printed in clay using an industrial robotic arm.
Silicon Valley challenger
A formidable rival has emerged in the United States — Meshy, a Silicon Valley-based AI 3D startup founded by Ethan Hu, an MIT PhD and an alumnus of Tsinghua University’s prestigious computer science talent program “Yao Class.” Hu previously gained recognition for creating the open-source graphics programming language Taichi.
At CES 2026, Meshy unveiled its AI Creative Lab, a platform built on a “prompt to product” philosophy, which can convert AI-generated 3D designs directly into full-color, ready-to-print physical objects like keychains and figurines, bypassing the need for complex computer-aided design software.
Meshy is also gaining traction among independent game developers. Stradion Studios utilized Meshy to generate a 3D tank model for a game in just four minutes from four hours of manual labor. Game developers are also using Meshy for “kit bashing,” a technique where artists combine fragments of pre-existing 3D models to rapidly construct entirely new, complex digital scenes.
Battling for dominance
Tripo and Meshy are slugging it out for supremacy, with both founders claiming their platforms are far more advanced than the competition.
Song Yachen, founder of VAST, says that after recent breakthroughs, his company’s technology maintains a “discontinuous, cliff-like lead” and that “there are basically no competitors in the market.”
Hu asserts that Meshy has the largest user base and traffic in the market. “Our product’s traffic is the sum of the second, third, and fourth places,” he said.
Despite the competing claims of technological superiority, a distinct pricing dynamic has emerged. Both have similar subscription prices but Meshy provides fewer credits for users to create 3D models. This reflects a broader trend in generative AI: China-based products tend to be cheaper and offer higher output volumes, while Silicon Valley-based products are often more expensive.
Looking ahead, both companies have ambitions beyond standalone tools. VAST says it is dedicating 2026 to the research and development of “world models” — AI systems that natively understand three-dimensional space and physics — with the goal of creating a massive, user-generated interactive content platform. Meshy, meanwhile, is betting on the democratization of manufacturing, hoping to bridge the gap between digital imagination and physical reality for everyday consumers.
Sources:
