Plurai raises $13m to test AI agents with self-driving car simulations

Key points

• Plurai raised $13m in seed funding led by Team8, with U&I, Velocity and Mercer participating.
• Co-founders Ilan Kadar and Elad Levi previously worked at Nexar, and Kadar helped Waymo simulate its driving software.
• Tests with Nvidia and Nebius showed Plurai's models beating mid-tier frontier models at lower cost, Kadar says.

Plurai, a startup that uses self-driving car simulation techniques to improve AI agents, has raised $13m in seed funding. Team8 led the round, with U&I, Velocity and Mercer taking part, TechCrunch reports. The company says its platform builds a digital twin of a customer's systems so agents can be tested against difficult edge cases before deployment.

The co-founders, Ilan Kadar and Elad Levi, both worked at Nexar, which develops autonomous driving technology. Kadar helped Waymo simulate and evaluate its end-to-end deep learning driving software. He told TechCrunch that building an agent demo is easy, but making one reliable and cost effective is hard, and that the same pattern led the team to build a simulation layer.

Plurai maps the possible flows an agent can take and uses that graph to prepare evaluations and fixes. It works alongside tools such as Claude Code, Codex, Copilot and Cursor. Kadar says the bigger promise is post-training small open-weight models on the data gathered, and tests with Nvidia and Nebius showed these models outperforming mid-tier frontier models at far lower cost.

Kadar argues that even much smarter frontier models will not know a specific business and its policies.

Read the full story on Techcrunch →