Why we invested in Vsim, a new physics simulation framework for robotics

The demand for faster, more accurate simulations is growing as both foundation model-based and reinforcement learning-driven robot training expand. With real-world data still scarce, simulation training is becoming critical to bridge this gap and accelerate development. In sectors such as robotics, gaming, media, and manufacturing, however, companies struggle with long simulation times and scalability issues, particularly when trying to execute simulations that require parallel computing capabilities. Current solutions can take weeks to complete, limiting the speed and efficiency of innovation.

We invested in Vsim because it addresses a critical bottleneck in multiple industries—slow, inflexible, and expensive physics simulations that are unable to fully leverage modern GPU architectures. Their physics simulation framework allows simulations to be completed in minutes rather than days, without compromising quality. This technology leverages new algorithms optimized for multi-core hardware, providing 100x faster performance at a fraction of the cost. While initially focusing on robotics training, the platform has broader applications across industries such as aerospace, gaming, and visual effects.

We recently participated in Vsim’s $21.5M funding round, led by EQT Ventures with additional backing from Tru Arrow and IQ Capital. This investment will fuel the growth of Vsim’s team and accelerate the development of their full-stack solution for revolutionizing robot training and simulation. With the robotics market, especially for general-purpose use cases, rapidly expanding, Vsim’s ability to train and fine-tune robot foundation models is crucial in speeding up product roadmaps and reducing costs.

We invested in Vsim not just for the technology but because of the founders, Michelle Lu and Kier Storey. Their years of collaboration, including 12 years at NVIDIA, where they built Isaac Sim and the world’s first GPU-accelerated robotics reinforcement learning framework, sets them apart. They first began working together over 20 years ago under the same Ph.D. supervisor, focusing on distributed physics simulation. Michelle and Kier’s contributions continue to power current systems, reflecting the lasting impact of their work. Prior to NVIDIA, they gained valuable experience at Activision's Bizarre Creations, where they developed a physics engine from scratch. Their technical expertise, combined with a proven ability to drive innovation across gaming and robotics, makes them the ideal team to lead Vsim’s growth and revolutionize simulation technology.


Niko Ciminelli is an investor at Samsung Next. Carlos Castellanos, another investor at Samsung Next, co-led this deal. Samsung Next’s investment strategy is limited to its own views and does not reflect the vision or strategy of any other Samsung business unit, including, but not limited to, Samsung Electronics.

Previous
Previous

Why we invested in Centaur Labs, the modern AI data labeling solution for healthcare

Next
Next

Why we invested in point.me, a real-time award travel search engine that simplifies booking trips with points and miles