CVC Series AI + Simulation: From Materials Innovation to Industrial Optimization

01/23/2026

CM Venture Capital’s CVC sessions are designed to give corporates and LPs targeted access to high-quality startups aligned with defined strategic priorities, offering early visibility into technologies with clear industrial relevance and investment potential.

The AI + Simulation CVC session assessed how simulation is transitioning from a niche R&D function to a scalable decision layer across the industrial value chain, enabled by five compounding capabilities:

  • Rapid GPU compute gains that increase throughput and reduce unit compute cost, making physics-grade simulation economically across field;
  • AI surrogate models that replace the most computationally expensive solver steps;
  • AI pattern discovery and reasoning that learns latent structure to support generation & reasoning;
  • AI pattern discovery and reasoning that learns latent structure to support generation & reasoning;
  • AI-enabled robotics enabling high-throughput labs that convert hypotheses into scalable experiments;
  • A closed-loop data flywheel where real-world measurements and outcomes continuously feed back into models to improve performance and expand coverage.

We invited startups across scales from microscale of materials discovery to macroscale of lab automation and clinical applications. A consistent theme across the session was practical deployment: moving beyond proof-of-concept toward solutions embedded in real-world workflows, where speed, accuracy, and integration with existing systems matter as much as model performance.

At quantum level, Kairos Materials uses a DFT-trained generative model to sample physically plausible crystal structures conditioned on target properties, then combines fast surrogate property prediction with selective DFT verification to run a generate–screen–validate loop for inverse materials design.

At formulation level, Polymerize, is a cloud-based materials informatics platform for chemical R&D that convert formulation recipes, process parameters, characterization and workflow into a unified dataset; on top of this, it applies AI to predict properties, recommend next experiments, and propose candidate formulations.

At the system level, the focus shifted to scaling simulation and automation across industrial and laboratory workflows. DynaFlow shared how its automated workstations and standardized data interfaces enable high-throughput experimentation with consistent protocols. The resulting structured data feeds and AI layer recommending next-best experiments, making experimentation more repeatable and data-driven.

Finally,BOEA Wisdom demonstrated how simulation is being embedded into clinical decision-making. The company’s software transforms CTA scans into 3D reconstructions and sizing measurements and is expanding into simulation-driven pre-operative planning. Its in-house FEM engine enables order-of-magnitude faster simulations, making near-real-time procedural guidance increasingly feasible.

The session combined CM’s research perspective with curated startup use cases and open discussion among corporate and LP participants, creating a shared view on where AI + Simulation is delivering value. Thank you to all speakers and participants for the thoughtful questions and candid exchange. We look forward to continuing the conversation in future CM CVC sessions.