Gravity / On-premises AI
Gravity installs and operates camera systems and GPU infrastructure for physical security, presence detection, and private transcription. Your video and audio are processed on site.
Founded by alumni of
What it does
Video, sensor, and speech models share the same on-site infrastructure. Enable the capabilities each space needs.
Physical security
Detect arrivals at the gate, activity at service entrances, and presence after hours. Define the zones and schedules that matter. Keep recordings on site for authorized review.
Video analysis · Event detection · Local recording
Presence & room controls
Understand which rooms are occupied without identifying every person. Connect presence to lighting and climate. Personal settings are available for people who choose to enroll.
Occupancy · Building controls · Personal settings
Private transcription
Turn meetings into text using speech models on your GPU server. Choose the rooms, recording permissions, and retention. Audio stays off cloud transcription services.
Local speech models · Recording permissions
Voice & A/V
“Make me comfortable.” Adjust the lights, climate, shades, or A/V through the controls already in the room.
How it’s built
Cameras, network, compute, and software are specified together. Gravity takes responsibility from installation through operation.
Around the property
Inside the rack
Loading the 3D room…
Drag to rotate · Use the controls to explore
GPU option / NVIDIA Blackwell
96 GB of GPU memory gives local video and speech models room to run. We select the GPU edition and server chassis around the workload, available power, and airflow.
NVIDIA hardware specificationsCompute infrastructure
Camera count, video resolution, model size, and recording retention determine the build. So do rack space, power, and cooling.
Gravity BrainDedicated GPU nodeCarries camera feeds and sensor data across the property.
Power and a wired connection for each compatible camera and sensor.
Connects microphones and speakers for enabled audio features.
Connects to the existing Lutron, Crestron, Savant, or other building controls.
Compute
Video analysis, transcription, and recording compete for resources. We plan GPU memory, storage, and network capacity together.
Installation
We install the server, switches, cameras, and sensors, and integrate equipment that fits the plan.
The GPU server runs video analysis and speech models locally.
Backup power and airflow are part of the rack design.
Video and audio are processed on the property. Hardware health reporting is separate from footage and recordings.
How we work with you
One team for camera placement, the server rack, and the software running on it.
New build
We work with your architect, integrator, and contractor on sightlines, cable runs, network capacity, and rack space before walls close.
Retrofit
We review the cameras, cabling, network, and storage already installed. Keep what works, replace what limits the system, and add the GPU server.
Rack, network, sensing, controls, and compute.
Coordinate trades and build the system.
Test camera coverage, detection rules, and access.
Monitoring, model updates, and hardware refresh.
When software controls a room
Detecting activity and controlling a door are different jobs. You decide which actions the system is allowed to request.
Video and sensor events are evaluated locally. Alerts and room controls follow the rules agreed during setup.
Lighting, climate, and access requests pass through the existing control system and its permissions.
Presence detection does not require naming a person. Personal recognition is available for people who choose to enroll.
Recording, recognition, and room controls are configured separately for each space.
Where we install
We plan around the people who use the property and the people responsible for it.
Luxury residences
Gates, driveways, deliveries, and occupied rooms. Video stays on the property; room controls fit the household.
Family offices & investment firms
Entrances, shared areas, and meeting rooms. Local video analysis and private transcription on hardware at your office.
Careers at Gravity
Our work spans GPU infrastructure, computer vision, speech models, and building controls. We’re looking for people who want to build and deploy the whole system.
Connect with our teamFind us
Coming soon
Talk to us
What cameras do you have? What do you need to see? Tell us about the site and we’ll start there.
