Honeywell has announced Forge Cognition, an operational AI architecture designed to bring autonomous software agents into commercial facilities, campus environments, and industrial complexes. The initiative represents an expansion of Honeywell's enterprise software platform, realized through a technical collaboration with Nvidia and Dell. The system aims to transition facility operations away from passive oversight toward real-time automated intervention. With real estate operators facing persistent pressures around utility spending and carbon compliance, the move signals an effort to link facility automation directly with high-performance computing.
At the architecture level, Forge Cognition pairs an agentic AI framework with digital simulation capabilities. Honeywell built the solution on Nvidia's Nemotron language models and Omniverse platform, allowing the system to model complex building physics alongside operational routines. Dell Technologies provides the computational backbone through its Dell AI Factory server infrastructure, which handles the localized processing demands of the models. By joining simulation engines with autonomous agent logic, the platform can evaluate operating scenarios virtually before deploying parameter adjustments to physical plant equipment.
In operational deployments, Forge Cognition ingests uninterrupted streams of live telemetry generated by building management systems. It interfaces directly with building controls, including heating, ventilation, and air conditioning (HVAC) systems. Autonomous AI agents continuously evaluate incoming sensor readings to monitor equipment performance, airflow, and ambient conditions. Unlike conventional rules engines that trigger based on static thresholds, these agents interpret anomalies within broader operational context, enabling early coordination among separate subsystems and identifying equipment drift before failures occur.
Initial operational figures from pilot deployments, including testing conducted at Nvidia's Voyager Campus, indicate significant improvements in control room efficiency. One of the primary operational bottlenecks in large properties, alarm fatigue caused by constant nuisance notifications, saw substantial mitigation. The pilot data showed a reduction in irrelevant alarm messages of up to 90 percent. By filtering out routine operational fluctuations, the agentic architecture frees engineering and maintenance teams to focus resources on genuine mechanical faults and urgent safety issues.
The pilot results also documented tangible improvements in utility efficiency. Through predictive and continuous adjustments of HVAC units and industrial equipment, Forge Cognition delivered savings of up to 10 percent in energy and utility expenses. The system throttles power consumption across unoccupied zones and aligns heating and cooling cycles with real-time occupancy and environmental variables. This proactive modulation avoids the energy waste associated with fixed schedules and softens costly peak-demand spikes across the facility grid.
The joint deployment by Honeywell, Nvidia, and Dell reflects a broader shift across smart facility management from passive monitoring toward autonomous execution. Traditional dashboards that merely present sensor histories are proving insufficient for managing increasingly complex commercial properties and industrial sites. As hardware vendors, industrial automation specialists, and model developers converge, operational AI agents are positioning themselves as a central control layer for commercial real estate.