Antioch raises $32M Series A to scale simulation platform for robotics

Sep 8, 2026

Antioch, a physical AI simulation startup founded in May 2025 by Harry Mellsop and others, raised a $32 million Series A funding round led by Greylock Partners on September 8, 2026, bringing its total funding to $44.75 million. The company's platform addresses the physical AI testing bottleneck by running high-fidelity, parallel simulations on cloud infrastructure. Amazon's Ring validates the technology, which also integrates Nvidia's open physical AI stack and partners with Nebius Group for cloud infrastructure.

Antioch Series A funding round

▪Prior to the Series A, Antioch raised a $4.25 million pre-seed round in December 2025 and an $8.5 million seed round in April 2026 at a $60 million valuation.
▪Individual angel investors in Antioch's Series A round included Palantir CTO Shyam Sankar, Foxglove CEO Adrian Macneil, and Nvidia executive Ian Andrews.
▪Antioch raised a $32 million Series A investment round led by Greylock Partners on September 8, 2026.
▪With the completion of the Series A round, Antioch has raised approximately $44.75 million in total funding.
▪Antioch's Series A round included participation from A* Capital, Category Ventures, BoxGroup, and Icehouse Ventures.

Cloud-based simulation platform capabilities

▪Antioch's platform generates synthetic sensor data, including radar, lidar, infrared, and non-visual sensors, to validate how systems respond to edge cases.
▪Antioch was founded in May 2025 by Harry Mellsop, Alex Langshur, Michael Calvey, and two former employees from Google DeepMind and Meta's Reality Labs.
▪Antioch's platform builds high-fidelity simulations for a customer's specific hardware and runs thousands of parallel evaluations on cloud infrastructure.
▪Antioch delivers simulation data across multiple verticals, including aerial autonomy, drone delivery, warehouse AMRs and AGVs, smart security systems, industrial automation, and construction robotics.

Amazon Ring customer validation

▪Amazon's Ring is using Antioch's simulation platform to improve drone and device development and reduce false positives in smart security systems.
▪Jason Mitura, vice president of software development at Amazon and chief product officer at Ring, stated that Antioch's simulations closely matched physical test results in scenarios held out of calibration.

Physical AI testing bottleneck

▪Physical AI development is bottlenecked because hardware and policy changes must be re-validated on physical equipment, which is slow and expensive.
▪Antioch's platform aims to address the 'sim-to-real gap,' which is the persistent mismatch between virtual environment performance and physical world performance.

Nvidia Omniverse platform integration

▪Antioch's integration of Nvidia's physical artificial intelligence stack is designed to help developers spread testing across thousands of simulated scenarios.
▪Antioch partnered with Nvidia to integrate Nvidia's open physical artificial intelligence stack, including Omniverse libraries, Isaac Sim, and Isaac Lab.

Nebius cloud infrastructure partnership

▪Antioch partnered with Nebius Group N.V. to supply the cloud infrastructure required to run Antioch's parallel simulation evaluations.
▪Evan Helda, head of physical AI at Nebius Group N.V., characterized simulation as one of the hardest parts of the physical artificial intelligence stack to get right.