Skip to main content

Video Rebirth

Category: Generative Media (Image / Video)

An industrial-grade AI video generation engine designed to provide cinematic quality, physical consistency, and high controllability for professional media production. Video Rebirth was founded in 2024. The company is led by Wei Liu. Based in Singapore, Singapore. Team size: 11-50. Total funding raised: $80.0M. Latest round: Series A (Mar 2025). Key investors include AMD Ventures, Hyundai Motor (ZER01NE), CJ Group, Qiming Venture Partners, ActozSoft.

Founded
2024
Headquarters
Singapore, Singapore
Team size
11-50
Total funding
$80.0M

Value proposition

Bridges the gap between AI-generated 'demos' and commercial-ready content by ensuring physical realism, temporal coherence, and precise frame-by-frame control.

Products and solutions

Bach Series (Frontier video generation models), Industrial-grade AI Video Engine, Physics-Native Attention (PNA) Architecture, Native 30fps Video Generation System, Dual DiT Prompt Adherence Engine

Unique value

Unlike consumer-grade tools that prioritize stylistic novelty, Video Rebirth focuses on 'World Models' that obey physical laws, eliminating visual artifacts and object distortion common in diffusion models.

Target customer

Professional film and TV studios, advertising agencies, e-commerce platforms, automotive designers, and AAA game developers.

Industries served

Media & Entertainment, Advertising & E-commerce, Automotive & Mobility, Gaming & Animation, Industrial Simulation

Technology advantage

Pioneered 'Physics-Native Attention' (PNA) to simulate spatio-temporal dynamics accurately. Led by Dr. Wei Liu (IEEE/AAAS Fellow), the company leverages strategic hardware optimization through its AMD partnership and industry-specific data from Hyundai and CJ Group.

How they differentiate

Focuses on 'Physics-Native Attention' (PNA) and 'Industrial-grade' output to ensure physical consistency and temporal coherence, moving beyond the stylistic 'hallucinations' of consumer-grade models to meet professional cinematic standards.

Main competitors

OpenAI (Sora), Runway (Gen-3 Alpha), Luma AI (Dream Machine), Kling AI

Key partnerships

AMD Ventures (Hardware and compute optimization), Hyundai Motor Group / ZER01NE (Mobility and simulation applications), CJ Group (Entertainment and content distribution), ActozSoft (Gaming and interactive media), Qiming Venture Partners (Strategic investment)

Notable customers

Hyundai Motor Group, CJ Group, ActozSoft

Major milestones

Secured $50M Series A funding in November 2025, topped up to $80M in March 2026, Developed proprietary 'Physics-Native Attention' (PNA) architecture, Established strategic hardware partnership with AMD Ventures, Launched BACH industrial-grade AI video engine in May 2026, BACH-1.0 Preview ranked #6 on Artificial Analysis Text-to-Video Leaderboard

Growth metrics

Launched BACH 1.0 Preview in May 2026; ranked #6 on Artificial Analysis Text-to-Video Leaderboard; full BACH 1.0 release forthcoming.

Market positioning

High-end enterprise generative AI provider for professional film studios, automotive designers, and AAA game developers.

Geographic focus

Global (Headquartered in Singapore with strategic focus on North America, South Korea, and China).

Patents and IP

Proprietary IP includes the Bach engine architecture, PNA logic, and MSSL (Multi-Scale Spatio-temporal Learning) technology; CEO Wei Liu holds extensive prior patents in computer vision and image-text matching.

About Wei Liu

Dr. Wei Liu is a world-renowned AI scientist and an IEEE, AAAS, and IAPR Fellow. He previously served as a Distinguished Scientist at Tencent, where he led large-scale AI research initiatives. His career includes significant tenures at IBM Research and Didi, and he holds a PhD from Columbia University, where he was advised by Professor Shih-fu Chang.

Official website: