Honeywell's 'Brain Behind the Building' Can Run Autonomously—It Won't, Yet

Honeywell Technologies launched Forge Cognition on Wednesday — an AI platform that can run buildings autonomously but won't, for at least two to three years, because its chief technology officer hasn't switched that mode on.

"We have not opened up an autonomy right now," Suresh Venkatarayalu, Honeywell's CTO, told Fortune ahead of the launch. "Maybe we will start running in this mode for probably next two to three years."

The platform, built with technology from Nvidia and Dell, puts AI agents and digital twins on top of the systems that heat, cool, secure and power buildings. Honeywell's launch announcement says the agents can "autonomously coordinate actions across existing systems." Venkatarayalu says people stay in the loop — for now. After the two-to-three-year window, customers decide when to make the switch.

"For me, I wouldn't deploy these agents without my control system in place," he said.

Venkatarayalu calls the platform "a brain behind a building." For now, that brain advises, and the operators decide. The gap between what the software can do and what Honeywell will let it do is the bet.

How did Nvidia become the first test case?

The project started with a challenge at CES in January. Nvidia's leadership heard Honeywell's pitch for autonomous buildings and asked the obvious follow-up: "Who is your lighthouse customer?" In other words, where's the real building that proves it works?

"We don't believe in just POCs," Venkatarayalu said, referring to proofs of concept. Honeywell CEO Vimal Kapur turned the question around and asked Nvidia: "Why can't you be my lighthouse customer?"

Nvidia took the deal. Within 30 days, Venkatarayalu was in Santa Clara pitching Nvidia's operations and real estate team. The two companies then set a 90-day target: show meaningful results at Voyager, Nvidia's roughly 750,000-square-foot headquarters, by Honeywell's investor day in June.

The teams agreed on five performance targets, measured the baseline, and tracked progress weekly across about 900 pieces of equipment. By late April, Honeywell had enough to show investors a three-minute live demonstration.

The pace went from pitch to investor demo in roughly four months. That's the acceleration a lighthouse customer buys.

What did the AI actually fix?

Two things: alarms and energy.

Building systems err on the side of caution, so they flag everything. In Voyager's security and access-control system, close to 90% to 95% of alarms were false, and Nvidia had eight or nine people per shift clearing them, Venkatarayalu said. Honeywell took Nvidia's open-source Nemotron model, trained it on Honeywell's own engineering knowledge, and cut unnecessary alerts tied to the Building Management System by up to 90%, according to Honeywell.

Operators can check every call the agents make — open each alarm the system classified and mark whether they agree. "That's all about bringing in an operator in the loop," Venkatarayalu said, so the operator can learn to trust it.

On energy, the agents broke Voyager's utility bill down by floor, zone, and peak and off-peak hours, then flagged a heating pattern. The building was heating zones to "something like 58 degrees throughout the year," Venkatarayalu said, without sensing outside temperatures above that mark. "Why should I be running the heating system?"

"We have identified close to 10% of energy efficiency," he said. That's potential savings identified through smarter optimization of operating conditions — not yet achieved.

One figure is delivered, the other is projected. Both come from the same 90-day pilot.

How do you talk to a building?

In its simplest mode, Forge Cognition is a chatbot for infrastructure. "Can I talk to the building? Can I talk to a physical asset?" Venkatarayalu said. Operators ask about performance and trends, then click into a digital replica of Voyager — built with Nvidia's Omniverse simulation software — down to a single floor, zone or machine. Dell supplies the servers that run it on site.

Traditional building automation could "see and act": sensors collected data, operators responded. The new version adds "think" and "learn." AI models trained on Honeywell's installed base and engineering documents interpret what the sensors pick up, and operator feedback improves them over time.

"Autonomy for me is automation plus plus," Venkatarayalu said.

The learning varies by site. At Honeywell's Bangalore campus, the goal was energy savings. At Voyager, Nvidia's leadership told him: "For me, comfort is more important." Before any customer moves to full autonomy, he said, the system should have "10 out of 10 accuracy all the time."

Is this replacing operators?

"At least I do not think as replacement," Venkatarayalu said. "In an industrial environment, I need a co-pilot that aids and directs me and helps me to do my job better." In the U.S., he sees a labor shortage. Elsewhere, a shortage of skilled operators.

Nvidia's head of operations and real estate, Scott Ekman, described the same problem, Venkatarayalu said: the system now performs on par with his best-trained building operators, a team he spent 10 to 15 years building. As Nvidia adds more large, complex buildings, Ekman told him, it won't be able to hire and train that talent fast enough.

Predicting equipment failures matters for the same reason — some replacement parts take six to 18 months to arrive. Nvidia will also sign a service contract only after Honeywell's forward-deployed engineers train its operations team and chief engineer to keep feeding the models.

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

Why this launch matters now

The launch comes four months after Honeywell finished dismantling its conglomerate. Advanced materials became Solstice Advanced Materials in October 2025; aerospace began trading separately in June 2026. The breakup followed a portfolio review and pressure from activist investor Elliott Management. The remaining company sells automation for buildings, factories and process plants.

Honeywell says its portfolio spans 350,000 sites — the market Forge Cognition could reach. The company hasn't said how many will adopt it.

By August, the Voyager project had reached "a critical point," Venkatarayalu said. The next step is converting Nvidia into a paying customer: "It's no more a pilot or a POC." Nvidia is selecting more buildings — possibly two or three more in the U.S. and one in Berlin, though nothing is final. Honeywell is also piloting the platform at the Fairmont Hotel in Dubai, NEX Mall in Singapore, and its own campuses in Bangalore and Richardson, Texas, and has picked 12 test sites across 12 industries over the next six months, including a customer's plant operations in Nigeria.

The autonomous mode stays off until those customers decide to turn it on — and until it hits "10 out of 10 accuracy all the time."