FUNDING · STARTUPS · #3426
Plurai raises $13M seed to simulate company systems and harden AI agents
Plurai, founded by ex-Nexar engineers Ilan Kadar and Elad Levi, raised $13 million in a seed round led by Team8 to build a simulation layer that creates digital twins of customers' electronic systems and generates synthetic data to evaluate and post-train models for agent tasks. The startup says its platform can reveal edge-case failures, verify fixes, and post-train small open-weight models—tests with Nvidia and Nebius reportedly showed better-than-mid-tier frontier-model performance at lower cost.
KEY POINTS
- Plurai, founded by ex-Nexar engineers Ilan Kadar and Elad Levi, raised $13 million in a seed round led by Team8 to build a simulation layer that creates digital twins of customers' electronic systems and generates synthetic data to evaluate and post-train models for agent tasks.
- The startup says its platform can reveal edge-case failures, verify fixes, and post-train small open-weight models—tests with Nvidia and Nebius reportedly showed better-than-mid-tier frontier-model performance at lower cost.
- Simulation-based digital twins and synthetic-data post-training could materially improve agent reliability and reduce operational AI costs, making agent deployments more practical for enterprises.
WHY IT MATTERS
Simulation-based digital twins and synthetic-data post-training could materially improve agent reliability and reduce operational AI costs, making agent deployments more practical for enterprises.