Vast is China's newest AI unicorn — a three-year-old Beijing start-up that just raised close to $200 million to build AI that generates 3D models from a sentence.
On 1 June 2026, Bloomberg reported that Vast had raised nearly $200 million, pushing its valuation past $1 billion. The round was led by Ince Capital and a venture fund backed by China Life Insurance, with Genesis Capital joining existing investors Eminence Ventures and Primavera Venture Partners. Vast was founded in 2023 by Simon Song Yachen, a co-founder of the Chinese frontier lab MiniMax, and builds Tripo — also known as Tripo Studio — a system that turns text and image prompts into 3D assets. The raise follows a $50 million Series A in March 2026.
This is the newest name on a list that keeps growing: Chinese AI companies crossing into unicorn territory on the strength of a single, well-chosen modality. Vast picked 3D.
What happened
The numbers are clean. Vast raised close to $200 million, vaulting past a $1 billion valuation to unicorn status, per the Bloomberg report carried widely on 1 June. Coverage of the round named Ince Capital and a China Life-backed fund as leads, with Genesis Capital, Eminence Ventures and Primavera Venture Partners participating. The company is barely three years old.
Vast's product is Tripo, which generates 3D models — the textured, riggable objects used in games, films, e-commerce and simulation — from a text prompt or a single image. Reporting on the founder framed Vast as aiming to challenge Tencent and Google in 3D generation, a pointed ambition for a start-up its size. The throughline is talent: Simon Song built MiniMax, then left to build Vast.
💡The modality nobody is watching
Text, then images, then video — each modality minted its own wave of companies. 3D is the modality the public has barely noticed and the games, film and robotics industries cannot get enough of. Vast bet the next wave is dimensional.
Why 3D is the next modality
Generative AI has climbed a ladder of difficulty. Text came first, then images, then video. Each rung was harder than the last, and each one created an industry. 3D is the next rung, and it is steep: a 3D asset is not a flat picture but a geometry with surfaces, materials and structure that must hold together from every angle.
The demand is enormous and mostly hidden from the consumer eye. Every game needs thousands of 3D assets. Every animated film, every AR or VR scene, every e-commerce product viewer, every robotics simulation runs on 3D content that today is slow and expensive to make by hand. A system that produces usable 3D from a sentence collapses a cost that has throttled whole industries. That is the market Vast is selling into, and it is why a three-year-old company can be worth a billion dollars.
The talent diaspora is the pattern
The detail I keep circling is the founder. Simon Song helped build MiniMax, one of China's frontier labs, then left to start Vast. That move — frontier-lab veteran departs to found a focused, vertical company — is becoming the engine of China's AI sector, the way the PayPal Mafia once seeded Silicon Valley.
The frontier labs are becoming academies. People learn how to build intelligence inside them, then leave to point that skill at one problem each. The diaspora is the ecosystem.
— On frontier-lab talent flows, 1 June 2026 (https://www.scmp.com/tech/tech-trends/article/3323106/ex-minimax-co-founders-new-venture-aims-challenge-tencent-and-google-3d-ai-models)
I wrote about MiniMax's own M3 architecture recently; Vast is what happens next to the people who built it. The same flow runs through the wider Chinese sector I covered in the China-and-the-robots roundup — a deep bench of talent recombining into new companies faster than Western coverage can track. A $200 million round for 3D generation is small beside Anthropic's $30 billion, but it is a sharper signal: the capital is now funding the second and third layers of the stack, not just the foundation models.
China's app layer matures under pressure
Vast does not exist in a vacuum. It is rising at the exact moment the United States is tightening the chips China can buy. The two facts are connected. Cut off from the easiest path to ever-larger foundation models, Chinese capital and talent are flowing into the application layer — products that wring value from models that are good enough, rather than chasing the absolute frontier that export control has made expensive.
There is a forward link too. 3D generation is not just for games. World models and physical-AI systems — the kind NVIDIA just pushed with Cosmos 3 — are hungry for 3D scenes and simulated environments to train in. A company that can generate 3D worlds cheaply is sitting upstream of the robotics and autonomy boom. Vast looks like a 3D-asset company today. It may be a physical-AI infrastructure company tomorrow.
Frequently Asked Questions
These are the questions investors, developers and creators have been asking since Vast's raise was reported. Short answers follow, drawn from the Bloomberg-originated reporting and corroborating coverage.
What is Vast?
In short, Vast is a Beijing AI start-up, founded in 2023, that builds Tripo — a system generating 3D models from text and image prompts. The answer, simply put, is a generative-3D company that just became a unicorn. The key is the founder: data from the reporting shows Simon Song Yachen, a MiniMax co-founder, started Vast after helping build one of China's frontier labs.
How big is the Vast funding round?
Vast raised close to $200 million, crossing a $1 billion valuation. According to Bloomberg's report, the round was led by Ince Capital and a China Life-backed venture fund, with Genesis Capital, Eminence Ventures and Primavera Venture Partners participating. Data from the coverage shows it follows a $50 million Series A in March 2026 — a fast climb for a three-year-old company.
Why is 3D generation important?
Generative AI advanced from text to images to video, and 3D is the next, harder rung. According to the analysis, a 3D asset must hold together as geometry from every angle, not just look right as a flat image. The evidence for the opportunity is demand: games, film, AR, e-commerce and robotics simulation all run on 3D content that is slow and costly to make by hand.
Who is behind Vast?
Vast was founded by Simon Song Yachen, a co-founder of the Chinese frontier lab MiniMax. In other words, the analysis shows a familiar pattern — a frontier-lab veteran leaving to build a focused, vertical company. The investors span venture and insurance capital: Ince Capital, a China Life-backed fund, Genesis Capital, Eminence Ventures and Primavera Venture Partners.
What are the real risks for Vast?
Analysis of the company reveals three durable risks. First, competition: Vast is taking on Tencent and Google in 3D generation, far larger rivals. Second, the export-control squeeze on Chinese compute, which raises the cost of training. Third, a single-modality bet, leaving Vast exposed if 3D demand matures more slowly than the capital expects. Evidence shows the upside — physical-AI demand — is real but not yet proven at scale.
Vast is a small story with a large shape behind it. A frontier-lab veteran leaves, picks one hard modality, and builds a billion-dollar company in three years — under export controls that were supposed to slow China down. The capital has moved past the foundation models to the layer above, and 3D generation turns out to sit upstream of the robotics and world-model boom now arriving. The frontier is not only getting smarter. It is getting wider, and a great deal of that widening is happening in places the Western AI conversation barely watches.
Related on humphreytheodore.com:
