Honeywell Forge Cognition Brings Physical AI to Buildings But Keeps Humans in Control

Honeywell Launches Physical AI Platform With Nvidia and Dell Partnership

Honeywell Technologies unveiled Forge Cognition, a self-contained agentic AI and simulation platform designed to transform facility operations across buildings, campuses and industrial environments1

1

2

2

. The AI-driven platform for building automation, developed through the Honeywell, Nvidia and Dell partnership, combines real-time digital twins with autonomous AI agents to help operators manage heating and cooling systems, security systems, and power infrastructure3

3

. This marks the first major product launch from the newly streamlined Honeywell Technologies, which recently spun off its aerospace and advanced-materials divisions to focus on building automation and industrial operations1

1

.

Human Oversight Remains Despite Autonomous Capabilities

While Forge Cognition can operate in three modes—manual, semi-autonomous mode, and fully autonomous—Suresh Venkatarayalu, Honeywell's Chief Technology Officer, confirmed the company has not activated autonomous operation1

1

. "We have not opened up an autonomy right now," Venkatarayalu told Fortune, indicating operators will maintain human oversight for approximately two to three years before customers decide whether to enable full automation1

1

. "For me, I wouldn't deploy these agents without my control system in place," he emphasized, highlighting the critical balance between Physical AI capabilities and operational safety in facility operations1

1

.

Nvidia Voyager Campus Proves Platform Effectiveness

The platform's development accelerated after a January CES challenge when Nvidia asked Honeywell to identify a lighthouse customer1

1

. Honeywell CEO Vimal Kapur turned the question around, asking Nvidia to serve as the proving ground. Within 90 days, Honeywell demonstrated measurable results at the Nvidia Voyager campus, a 750,000-square-foot headquarters building in Santa Clara, California1

1

3

3

. The pilot tracked progress across approximately 900 pieces of equipment, focusing on five specific performance targets1

1

.

Smarter Alarm Management Reduces False Alerts by 90%

One breakthrough application involves smarter alarm management in building systems2

2

. At the Nvidia Voyager campus, security and access-control systems generated 90% to 95% false alarms, requiring eight or nine staff members per shift to clear them1

1

. Honeywell trained Nvidia's open-source Nemotron model on its engineering knowledge to distinguish genuine alerts from noise, achieving up to a 90% reduction in unnecessary alerts related to the Building Management System1

1

3

3

. Operators can review each classification decision and provide feedback, creating a learning loop that builds trust in the AI agents1

1

.

Energy Intelligence Identifies 10% Efficiency Opportunity

The platform's energy intelligence capabilities analyzed Voyager's utility consumption by floor, zone, and peak versus off-peak hours1

1

. AI agents discovered heating systems running at approximately 58 degrees throughout the year without sensing external temperature changes, questioning why heating operated when unnecessary1

1

. Honeywell identified close to 10% potential energy efficiency improvements through intelligent optimization of operating conditions, though these savings represent identified opportunities rather than achieved results1

1

2

2

.

Technology Stack Combines Nvidia Omniverse and Dell AI Factory

Forge Cognition's architecture integrates multiple technologies from the Honeywell, Nvidia and Dell partnership4

4

. Nvidia Omniverse libraries enable physically accurate simulations, creating a digital replica of facilities down to individual floors, zones, or machines1

1

2

2

. The Dell AI Factory provides underlying infrastructure through Dell PowerEdge XE servers, Dell AI Data Platform data and storage engines, and cyber resiliency solutions2

2

3

3

. The platform uses NVIDIA RTX PRO GPUs and is expanding to accommodate far-edge deployments with NVIDIA DGX Spark2

2

.

Chatbot Interface Enables Natural Building Interaction

At its simplest level, Forge Cognition functions as a chatbot interface for infrastructure, allowing operators to ask questions about performance and trends1

1

. "Can I talk to the building? Can I talk to a physical asset?" Venkatarayalu explained, describing how users can query systems and then drill into specific components within the digital replica1

1

. Traditional building automation could "see and act" through sensors and operator responses, but the new platform adds "think" and "learn" capabilities by training AI models on Honeywell's installed base and engineering documents1

1

.

Predictive Asset Health and Early Deployment Sites

Beyond alarm management and energy optimization, the platform offers predictive asset health monitoring by analyzing real-time HVAC data to identify emerging equipment issues before they disrupt operations2

2

3

3

. Honeywell is piloting the agentic AI and simulation platform at multiple locations including Fairmont Hotel Dubai, NEX Mall Singapore, Vanderbilt University, and its own campuses in Bangalore and Richardson, Texas3

3

4

4

5

5

. These deployments provide early examples across Honeywell's install base of more than 300,000 sites3

3

4

4

. As Deepu Talla, Nvidia's vice president of robotics and edge AI, noted, "Physical AI is bringing new intelligence to the complex systems that run our buildings and industrial facilities"2

2

.