Antioch has raised $8.5 million in seed funding to develop simulation tools for robotics and physical AI systems, targeting a key bottleneck in the industry known as the “sim-to-real” gap. The round, which values the company at $60 million, was led by A* and Category Ventures, with participation from MaC Venture Capital and others.
Founded in 2025, the New York-based startup is building software that allows developers to train and test robots in high-fidelity virtual environments. The goal is to reduce reliance on costly physical testing setups, such as mock warehouses or sensor-heavy real-world trials, which currently limit scalability in robotics development.
Antioch’s platform enables developers to create digital replicas of hardware systems and simulate real-world sensor data, allowing for tasks such as reinforcement learning, edge-case testing, and data generation. The company builds on foundational models from providers like Nvidia and other AI labs, layering domain-specific tools to improve usability and realism.
The approach reflects a broader shift toward simulation-driven development across autonomy sectors. Companies like Waymo already rely heavily on virtual testing to scale self-driving systems, highlighting the potential for similar methods in robotics.
As demand grows for physical AI applications in areas such as logistics, agriculture, and robotics, Antioch is positioning itself as a core infrastructure provider. The company argues that better simulation tools could enable faster iteration cycles and lower costs, helping more companies build and deploy autonomous systems without massive upfront investment.
