Antioch has raised $32M in a Series A led by Greylock. The company builds a simulation platform for physical AI — it calibrates high-fidelity simulations to a customer's specific hardware and runs them at cloud scale. The round was reported on 9 September 2026.
Key takeaways
- $32M Series A, led by Greylock
- Co-investors: A*, Category Ventures, Box Group and Icehouse Ventures
- Angels: Shyam Sankar (Palantir CTO) and Adrian Macneil (Foxglove CEO)
- The only named customer is Amazon's Ring
- Greylock lists the company as Deep Tech at Series A stage
What Antioch sells
The problem is mundane. Robotics teams test on physical hardware they do not have enough of, under conditions they cannot reproduce.
Antioch moves part of that cycle into simulation. The load-bearing word is "calibrates" — the simulation is meant to match a customer’s specific hardware unit, not a generic model. Running in the cloud lets teams multiply passes without growing the robot fleet.
Who is behind the round
Greylock leads, and the fund's portfolio lists Antioch as Deep Tech at Series A stage. Saam Motamedi owns the investment. NVIDIA and Nebius are technology partners.
The angel list is more telling. Shyam Sankar is Palantir's chief technology officer, and Adrian Macneil runs Foxglove, a robotics data visualisation tooling company. These are operators, not purely financial backers.
What was not disclosed
No valuation, no customer count, no technical detail. It is not known which physics engine underpins the platform, nor how simulation fidelity is measured.
The only named customer is Amazon's Ring — home devices, not humanoids or industrial robots.
Why it matters
Simulation is a bottleneck discussed far less than models. Every control policy needs thousands of passes, and a pass on a physical robot costs time and wears the hardware out. Calibrating to a specific unit is a sensible answer to the gap between simulation and reality. The condition is proof that the transfer holds, and the company has published no figures comparing results in simulation against hardware.
What's next
- The round is meant to fund broader test coverage without growing the physical robot fleet
- No customer beyond Amazon's Ring has been named, making that the first traction metric to check
- The company has announced no plans to publish simulation-to-hardware transfer benchmarks
Sources
- The AI Insider — Antioch Raises $32M in Series A Funding for Physical AI Simulation Platform
- Greylock — Portfolio
