.png)
Stanford is stacking BMS sensors, scheduling systems, space-planning databases, and occupant input to hit the confidence threshold labs demand for unoccupied airflow cutback
Stanford is looking to expand its occupancy-driven operations to laboratories. Given the criticality of these spaces, no single sensor delivers the confidence threshold required to adjust the space's HVAC.
Labs are the highest-energy spaces on a campus like Stanford's, making them the biggest operational prize for cutting airflow when no one is in them. They are also the most dangerous place to make a wrong call. An unoccupied teaching lab setback saves real money. Losing experiments due to temperature or humidity excursions can wipe out months of research and the grant funding that supports it.
"It's not just a sensor, it's not just binary. It's what level of confidence do you have that nobody is using the space," said Gerry Hamilton of Stanford, who manages the university's smart building and automation systems, at NexusCon 2025. "How many nines of confidence can you build into this?"
The team is stacking four proxy data sources: the HVAC system's existing occupancy signal (typically room sensors and CO2 inputs), room-scheduling systems, space-planning databases that document who works in a space and what they typically do there, and direct occupant input through a custom-built web portal.
On top of the proxy stack, Hamilton sees AI as the interpretive layer. It reads across the sources to conclude that no single one could deliver alone.
The Stanford team is approaching occupancy-driven operations in laboratories with caution, understanding that trust in the program can be easily lost if decisions aren't made with sufficient confidence.
Register for the next Nexus Labs event.
Sign up for the newsletter to get 5 stories like this per week:
Stanford is looking to expand its occupancy-driven operations to laboratories. Given the criticality of these spaces, no single sensor delivers the confidence threshold required to adjust the space's HVAC.
Labs are the highest-energy spaces on a campus like Stanford's, making them the biggest operational prize for cutting airflow when no one is in them. They are also the most dangerous place to make a wrong call. An unoccupied teaching lab setback saves real money. Losing experiments due to temperature or humidity excursions can wipe out months of research and the grant funding that supports it.
"It's not just a sensor, it's not just binary. It's what level of confidence do you have that nobody is using the space," said Gerry Hamilton of Stanford, who manages the university's smart building and automation systems, at NexusCon 2025. "How many nines of confidence can you build into this?"
The team is stacking four proxy data sources: the HVAC system's existing occupancy signal (typically room sensors and CO2 inputs), room-scheduling systems, space-planning databases that document who works in a space and what they typically do there, and direct occupant input through a custom-built web portal.
On top of the proxy stack, Hamilton sees AI as the interpretive layer. It reads across the sources to conclude that no single one could deliver alone.
The Stanford team is approaching occupancy-driven operations in laboratories with caution, understanding that trust in the program can be easily lost if decisions aren't made with sufficient confidence.
Register for the next Nexus Labs event.
Sign up for the newsletter to get 5 stories like this per week:
This is a great piece!
I agree.