RBC Global Industrials Conference keynote, November 17, 2026: Chairman and CEO Vimal Kapur outlines the newly independent automation company

Honeywell Chairman and CEO Vimal Kapur used the opening keynote of Day 2 at RBC's Global Industrials Conference to describe the company that remains after two spin-offs, three divestitures and six acquisitions. In a fireside chat with CNBC anchor Brian Sullivan, Kapur said Honeywell Technologies is now a pure-play automation company focused on mission-critical environments. He said data center orders are up 100% year on year, that LNG equipment bookings already extend to 2029, and that physical AI is the main long-term growth driver.

Data Center Orders Doubled, With Demand Broadening Beyond the U.S.

Kapur said Honeywell's data center business, which covers safety, security and energy management automation, is seeing demand that shows no sign of slowing. "Our base is small, but orders are up 100% year-on-year," he said, referring to what the company disclosed on its third-quarter earnings call. He added that the 2027 forecast is also "very robust."

Kapur argued that the market is underestimating data center growth outside the U.S. He pointed to demand in Europe, East Asia, Japan, Korea and India, and to the Middle East's ambition to become a data center hub. In his view, any U.S. restrictions or local pushback on data centers would be offset by demand elsewhere. "Some offset here, any demand reduction, if any, will get easily offset by the demand increasing in other parts of the world," he said. Sullivan, who said utility executives he spoke with last week saw no slowdown, raised the point about U.S. regulatory risk. Kapur's answer rested on the global footprint.

LNG: A 30-Month Build Cycle Gives Visibility to 2030

Kapur described Process Automation Technology as a segment of about $8 billion, roughly two-thirds of which is tied to the energy sector, including refineries, petrochemical plants, LNG facilities and renewable fuels. He said LNG is currently the largest area of uptake. Honeywell supplies the technology that liquefies gas, and Kapur said that "all big customers in the United States and Qatar" use it, although he noted the company does not have 100% share.

He described Honeywell's content in a typical LNG facility as three components. The first is front-end gas cleaning to remove impurities. The second is the heat exchanger, the tall towers that convert gas to liquid. The third is the control system in the control room, which manages safe production and throughput. Kapur said the lead time makes the business a leading indicator. "It takes more than 2 years to build one of our equipment, almost 30 months. So if you need something in 2028, you need to commit to me now." He said bookings already extend to 2028 and 2029, that the cycle should stay strong through 2027, and that the company could remain busy with shipments and plant set-ups "up to 2030."

Kapur said a customer he met the day before, who is considering another facility, attributed the boom to two factors. One is general growth in power demand, which requires gas because few regions will build new coal plants. The other is the conversion of retiring coal plants to gas. Kapur said he is "not a specialist" who can predict when the cycle will peak, but that demand for Honeywell's LNG capability "remains very high across the board." He said the LNG up-cycle has run for at least three years.

Physical AI: A Learning Layer on Top of Automation, With Humans Still in the Loop

Kapur defined physical AI as much broader than robotics or humanoids, which he said Honeywell does not pursue. He traced the logic from automation's origins in the 1970s, when computers took over from pneumatic and hydraulic systems. Automation handled the complex logic and humans managed exceptions. Over time, he said, facilities became more complex as capabilities and third-party systems were added, and human operators could no longer keep up. Physical AI, in his description, supplements the control system "with a learning system, which runs along with it, to tell humans what may happen if the current situations are prevailing, so that humans have more tools at hand to run the facilities better."

He said the human stays in the loop and that the case for the technology is partly demographic, as retirements and skill shortages make facilities harder to run. He also said customers buy it for revenue, not cost. "Our customers don't get attracted if we say we're going to use AI. But this tool is for them to drive uptime or asset throughput, and that is really what attracts them far more." He conceded that semiautonomous systems may reduce headcount, saying "probably yes," but noted that the industry is short of labor anyway.

Kapur was direct about why adoption is hard. Unlike consumer AI, the data is not public. It sits in proprietary systems at Honeywell and its peers, and it must be trained for each domain, because a hotel, a hospital and a refinery each care about different things. Honeywell uses large language models, but he said building the ontology layer, which establishes the relationships among different assets, is critical. He compared the timeline to driverless cars. "When did the driverless car start the journey? 2009. Here we are, 17 years later, and it works." He said physical AI is already being integrated into Honeywell's offering and will scale "every quarter, every year."

Corporate Structure, Naming and the $12 EPS Target

Kapur said the company completed a two-year restructuring. Advanced Materials has been listed since October 2025 and Honeywell Aerospace was spun off at the end of June. Honeywell Technologies reported its first standalone quarter in the third quarter, and 2027 will be its first full year. Automation represents more than 80% of the business, but the company chose not to call itself Honeywell Automation. "We don't know where this company will go over the next 50 years, so we wanted to have the freedom of practice to leave the future generation to imagine where to go with the company."...