PassiveLogic's CEO says one person can run 100 buildings with its AI

One person, a home office, a bank of screens, 100 buildings. That's how Thomas Kiessling, CEO of Utah-based PassiveLogic, described the company's operating model to Utah Business reporter Catherine Bennett in an interview published Oct. 29. He compared it to Waymo, where the person overseeing a car can be hundreds of miles away.

Kiessling was firm that the system doesn't replace people.

A ratio of one supervisor to 100 buildings is a staffing plan as much as an energy plan, and it lands hardest on portfolios that can't fill the jobs they already have. Whether it holds up depends on what sits underneath it: the controller, where it has been deployed and how large the savings really are.

From DOE research project to building brain

PassiveLogic started about a decade ago as a Department of Energy-funded research project on digital twins. It now employs about 130 people, most of them in Utah, and holds more than 200 patents. Kiessling, a German-born serial entrepreneur with close to 30 years in industrial and clean tech and a Ph.D. in operations research, joined as a consultant a year before the interview and became CEO in May 2026. Asked about the hardest part of commercializing the technology, he gave a one-word answer: focus. The company was building a whole platform, he said, not a single invention.

Kiessling puts it more simply. Most buildings, he said, are blind: they don't know how much energy they use or how well their equipment is working. PassiveLogic's controller connects to every sensor, light, shade and ventilation unit, learns how the building behaves and keeps tuning for the largest savings it can find.

The same physics model is what makes it useful to the utility. Kiessling said the controller can pre-cool or pre-heat ahead of the morning and afternoon peaks the grid has to carry. By his account, flattening those peaks would save billions in low- and medium-voltage grid costs.

Why Europe bought first

Most of PassiveLogic's customers are in Europe, and Kiessling credits regulation. Some European countries move faster on climate rules, which he said made selling there easy. The largest autonomous building the company knows of was just commissioned in Amsterdam, with 392 PassiveLogic devices running it fully autonomously.

Prologis is the partnership Kiessling pointed to. European energy rules required the logistics landlord to put intelligent systems in its warehouses, and PassiveLogic's cost position made it an easy fit, he said. For owners with European assets facing similar rules, that turns the purchase into a compliance line item rather than a discretionary efficiency project.

Taken together, those figures indicate who the 100-building ratio is meant for. A US owner with no automation mandate can't fall back on compliance to justify the spend. The business case rests on energy and on covering a thinning bench of building engineers, which makes the savings numbers worth checking carefully.

Three savings numbers, three baselines

Kiessling's estimate is the most conservative. He attributes 40% of human carbon emissions to buildings and said that running them on PassiveLogic from the start would probably save 10-15% of that share. The company's newsroom post goes further, claiming up to 30% lower building energy and emissions from retrofitting autonomous control onto existing equipment, with nothing replaced. The same post cites a Department of Energy estimate that high-performance controls cut annual HVAC energy by about 30% on average, which it says would equal more than 3% of total US energy use if deployed broadly. It also cites European Commission research finding smart building management can reduce heating energy by up to 30% and cooling energy by up to 73%.

These figures don't contradict each other. They measure different things: a share of emissions, whole-building energy, HVAC energy alone, and heating and cooling separately. Even the denominator moves. The newsroom post, citing the UN Environment Programme, puts buildings and construction at more than 30% of global energy use and emissions, while Kiessling uses 40%. None of the sources say which building conditions produce 10% and which produce 30%.

Autonomy versus the work order

Some in facilities management doubt the autonomy pitch. Writing in Buildings in September, Zac Wolf argued that most facilities teams are adding AI on top of the work they already do and are not replacing their systems. In his view, facilities management depends on people: only a human can tell a routine repair from a failure that threatens the business, and AI works as an assistant, not an operator. His example was a national retailer whose store teams kept picking the wrong part types for lighting repairs. AI that suggested the right inputs as each request was created cut cancellations and made dispatch predictable.

The two views overlap more than they first seem to. Wolf is writing about work orders and dispatch, where the context changes every time. PassiveLogic's autonomy sits in the control loop, where physics sets the rules, and Kiessling's model still puts a person in charge of oversight. Read together, they suggest a split: machines run the sequences that physics can define, and people make the judgment calls. Wolf's warning about data also applies to both. AI can only find patterns that accurate asset, location and work-order records let it see.

Labor economists offer a framework for this question. In a February 2026 Hamilton Project paper, MIT's Daron Acemoglu, David Autor and Simon Johnson sort technologies into five categories and argue that only those that create new tasks are clearly pro-worker, because they create demand for new expertise instead of commoditizing existing skills. By that measure, the remote oversight role Kiessling describes could count as a new kind of expert role if the job builds deeper systems knowledge, or as plain automation if it just spreads the same skills thinner.

Next up: data center cooling with NVIDIA

PassiveLogic is now aiming the same controller at data centers. Kiessling said the company is ready to start development and is working directly with NVIDIA and its partners on a system to manage data center power and cooling. He called it the company's sweet spot, because the work comes down to thermodynamics and airflow.

The controller stays the same, he said, but the use cases are more real-time and more mission-critical than in an office or a warehouse. The work is at the development stage. For data center operators, the open question is whether a controller built to save energy in offices and warehouses can meet those real-time, mission-critical demands for power management and cooling.