Tripolar Industries, Inc.

Mr. Firas Khalifeh, Chief Executive Officer, Tripolar Industries, Inc.

Firas Khalifeh is the Founder and CEO of Tripolar Industries, a Physical AI company based in Dubai building autonomous operations for industrial facilities. Tripolar's platform runs on-premises at the edge and takes over control of facility infrastructure, including HVAC, chiller plants and thermal systems, across fabs, data centers and industrial plants, without operational data leaving the site. Tripolar is an NVIDIA industrial AI partner in the Gulf region.

Before Tripolar, Firas founded Carbon Mobile in Germany, where he invented HyRECM, a patented carbon fiber composite technology for electronics, and built the first consumer electronics assembly line established in Germany in decades. The work received the JEC Composites Innovation Award, and he was named to MIT Technology Review's Innovators Under 35.

He also holds an Venture Partner position with Omni Ventures, a manufacturing technology venture fund.

 

 

Topic:

The Future of Autonomous Facilities in Time of Sovereignty Need

 

Abstract:

Semiconductor manufacturing is expanding into new jurisdictions faster than the capability to operate it. Every new fab, advanced packaging line and AI data center commits its owner to decades of facility operation: cleanroom air handling, chilled water, exhaust and abatement, ultrapure water and power distribution. These systems drive the majority of non-process energy in a fab, are a leading source of unplanned downtime and environmental excursions, and are still run by rule-based building management logic and a shrinking pool of experienced facility engineers.

Tripolar Industries builds the AI layer that operates this infrastructure autonomously. The platform runs on-premises at the edge, learns the thermal and load behaviour of each facility, and transfers that learning across sites as model artifacts, so no operational data leaves the plant. Our current pilots proven 20% in energy saving and 25% in co2 emissions reduction. 

This session argues that facility autonomy, not process automation, is the near-term constraint on fab and AI data center expansion; that sovereignty requirements make on-premises inference an architectural requirement rather than a compliance preference; and that a facility which can be operated autonomously becomes a controllable asset on the grid instead of a fixed load. It closes with what has to be true technically for autonomous facility operation to be trusted in a production fab.

 

Key Technologies Covered:

• Edge AI inference for facility infrastructure, deployed on-premises and single-tenant
• Autonomous control of cleanroom HVAC, chiller plants and air handling units
• Ultrafast Digital Twin: live demo. 
• Critical Infrastructures using MPC (Model Predictive Control) 
• Sovereign AI architecture: operational data never leaves the site