Manufacturing Technology Project Launches AI-Powered Platform to Make Metal 3D Printing Predictive

Manufacturing Technology Project (MTP), an American deep-tech startup, has launched AdditiveM — a physics-based process and performance simulation platform for metal additive manufacturing that aims to transform the industry from empirical trial-and-error to predictive engineering.

From Guesswork to Prediction

"What immediately fascinated me was that additive manufacturing is not just a fabrication method, but a fully coupled physical system in which heat transfer, fluid flow, phase transformation, residual stress, microstructure evolution, and structural performance are all tightly interconnected," said Hamed Hosseinzadeh, President & Tech Strategist at MTP. "At the same time, I observed that much of the industry still relied heavily on trial-and-error and empirical tuning, which is costly, time-consuming, and risky for qualification-critical applications."

What AdditiveM Does

The AdditiveM platform acts as a digital twin of the printing process, capturing the full thermomechanical history of a part and linking it to:

• Residual stress and distortion prediction
• Crack-prone zone identification
• Fatigue-relevant stress-strain history
• Microstructure evolution

The platform uses coupled multiphysics and multiscale modeling including high-fidelity transient thermal simulation of layer-by-layer melting, thermo-elasto-plastic mechanical modeling with temperature-dependent plasticity, and stochastic AI-assisted modules for uncertainty-aware prediction.

Bridging the Gap

Founded to bridge the gap between high-fidelity physics and real-world manufacturing practice, MTP's vision is to create a new generation of engineering tools that combine multiphysics simulation, Physics-Based AI, and digital twin technologies to enable predictive modeling, process optimization, and eventually autonomous control of manufacturing systems.

The platform currently focuses on laser-based Powder Bed Fusion (PBF) but is developed as a process-agnostic computational framework.