Nvidia is doubling down on opportunities to build robotics and other industrial AI applications with the launch of the Omniverse platform and, most recently, Mega, the Omniverse Blueprint framework that creates digital twins to drive these applications. . The company is also investing in digital twin startups to get the initiative off the ground.
Using AI and 3D technology, Taiwan's MetAI has developed a model that can transform CAD files into functional 3D environments within minutes and rapidly generate “SimReady” digital twins.
NVIDIA is currently backing MetAI in its first round of funding. The $4 million seed round marks the semiconductor giant's first investment in a Taiwanese startup. Other investors in this round include a mix of other strategic and financial investors, including Kenmec Mechanical Engineering, Solomo Technology, SparkLabs Taiwan, Addin Ventures, and Upstream Ventures.
The next wave of AI, known as generative physical AI, will rely on physically accurate simulated environments to train and validate robots used in autonomous systems and build operational AI before deployment. Masu. MetAI claims that the digital twins it helps create will be central to its efforts.
“Digital twins have long been seen as a barrier to entry into physical AI because they require months or even years of effort to develop,” said Daniel Yu, CEO and co-founder of MetAI. he said in an interview.
MetAI focuses on AI-powered digital twins customized for advanced semiconductor factories, smart warehouses, and automation. It also generates synthetic data within an AI-enabled digital twin environment.
Renton Hsu, co-founder of Yu and CTO of MetAI, has a background in 3D engineering and AI and first started working with digital twins while building enterprise AI software applications. Digital twins were used as a practical workaround in situations where clients lacked digital twins. Enough data to train the system. He then realized that the same could be applied to 3D systems and integrated 3D technology with AI to develop a synthetic AI and 3D solution, and Yu (who joined the startup with experience in digital transformation projects) and 3 I worked with Dave, my co-founder. Liu (COO) starts MetAI.
This breakthrough was enough to earn it first place in an AA competition run by Nvidia, and Hsu became the country's “Jetson AI Ambassador.”
MetAI's competitors range from large companies to small businesses that have built digital twin technology for manufacturing. These include Siemens Digital Industries, Dassault Systemes, Hexagon AB, Duality AI, and Intagles. There are a number of companies in the synthetic data space, including Sky Engine and Scale AI.
We believe that MetAI has a unique approach compared to all of these.
“Unlike competitors who prioritize operational efficiency and IoT integration, MetAI leverages generative models and AI-driven layouts to create digital twins designed for physical AI training and implementation into real-world operations. ,” Yu said. “This approach not only accelerates the creation of digital twins, but also enables them to be used directly in advanced automation systems such as robotics, bridging the gap between simulation and reality.”
MetAI differentiates itself by generating artificial data within an AI-enabled digital twin environment. Yu said this will allow users to generate synthetic data customized to specific operational requirements, making it easier to train and validate AI. “Instead of creating isolated datasets, MetAI builds dynamic virtual worlds (i.e., world simulators) – realistic virtual environments that behave exactly like the real world,” he said. Ta.
The two-year-old startup, which offers a wide range of products from vertical AI agents to digital twins, has a small number of customers and has already made money by partnering with companies in the manufacturing and automation industries. We expect to earn $3 this year. One project can reach $1 million, Yu said. Revenue will come from project-based revenue, product subscriptions and licensing fees from ongoing development, he added.
“The integration of MetAI and NVIDIA Omniverse represents a transformational advancement for industrial digital twins and physical AI in simulation,” Nico Caprez, corporate development manager at Nvidia, said in a statement. “Their ability to create a scalable environment for AI training has the potential to set new standards in industries ranging from manufacturing to robotics.”
In 2023, MetAI collaborated with Kenmec to create a digital twin for automated warehouses. MetAI claims that its technology has significantly reduced the time required for warehouse digital twin simulation from thousands of hours to just three minutes, resulting in significant savings in operational and validation tasks.
With this funding, MetAI plans to expand its research and development team to accelerate the development and execution of go-to-market strategies to meet growing demand. In addition, the Taiwan-based startup plans to open a US office and relocate its headquarters in late 2025, Yu told TechCrunch.
“Taiwan serves as our testing ground, collaborating with Taiwanese industry leaders to integrate deep vertical knowledge into our models and ensure the robustness and scalability of our solutions,” said Liu. said. “Given its size and demand for simulation-driven solutions, high labor costs and operational complexity, we are expanding the U.S. market. Our expansion strategy includes both point and end-to-end solutions. This includes SaaS products and vertical AI agents designed for rapid implementation in real-world scenarios in these industries.