SKW EnergyOSby Kaivalya Innovations

Physical AI for buildings

Self-driving buildings.

If a car can drive itself, a building can run itself.

Physical AI that autonomously operates HVAC — perceiving, deciding, and acting on real equipment. Proven in production.

Perception · liveActive
Booting perception…
AHU-0124.1°
AHU-0224.0°
AHU-0324.3°
AHU-0424.1°
68 AHUs autonomous · safety bounds enforced
Proven in production

A real building has run itself for a year.

Not a pilot. Not a dashboard. A building operating its own HVAC, continuously, within hard safety bounds.

0+months

Continuous autonomous operation

0incidents

Comfort or safety incidents

0AHUs

Under autonomous control

0ksq ft

Operating itself, live

Live across portfolios including Brookfield, Johnson Controls, and Westin.

The shift

Everyone else advises. We operate.

Every other system stops at a recommendation. Ours runs the equipment — inside bounds it can never cross.

Five levels of building control

L4
Autonomous controlUs

The building runs itself.

L3
Advisory AI

Analytics that recommend.

L2
Rules-based BMS

If-this-then-that logic.

L1
Fixed schedules

Timers and static setpoints.

L0
Manual

Operators turn the knobs.

Advisory AI & analytics

Recommends

SKW EnergyOS

Operates

Who acts on the finding
Your operators, manually
The system, autonomously
Output
A dashboard and alerts
Setpoint commands on real equipment
Response time
Hours to days
Seconds, continuously
When conditions change
Waits for a human
Re-decides on its own
Improves over time
Static rules
Re-learns on your building
Safety
N/A — it only suggests
Every command bounded; human override always wins
How it works

One control loop.

Perceive. Decide. Act. Re-learn — continuously, on real equipment.

digital twin · live deploymentPerceive
Live scan · 312 points over BACnet
Sensor sweep · 01/04

Safety layer — every command is bounded before it is issued; the system can never harm the equipment, and it keeps operating safely even offline.

Under the hood

One loop. Four systems.

The autonomy is not a single model — it is four systems working as one closed loop.

Reads the building

Telemetry layer

Connects to your existing BMS over BACnet and Modbus, read-only to begin with. Streams every controllable point — no new panel, no open inbound ports, AMC contracts untouched.

  • BACnet / Modbus
  • Read-only onboarding
  • Works with Honeywell, JCI, Siemens, Schneider
Telemetry layer
Reads the building
AHU-01PRESSZONE-4BOATBMS
Live

A building operating itself, right now.

This is the deployment — AHUs under autonomous control, setpoints being written live, the safety layer green.

candor-techspace · live operations
AI controller: Active
Under autonomous control68/68
AHU-01
24.1°C
AHU-02
24.3°C
AHU-03
24.0°C
AHU-04
24.4°C
AHU-05
24.2°C

cycling · all 68 autonomous

Plant power draw · kW318 kW
24.0°C
Supply air
142 Pa
Static press.
24.0°C
Zone 4B
Controller
Autonomous

uptime 12m 04d · last cmd 1s ago

Safety bounds OK
Override ready
Energy vs. baseline today: −19.3% (secondary — an outcome of autonomy, not the goal)
How we prove it

Autonomy is earned, in stages.

Control is never handed over on day one. It is transferred gradually, only after the system has proven itself on your building.

Read-only
OBSERVE0 commandsno write path
01Read-only

Read-only shadow

Weeks 1–6: monitors only, issues no commands. Findings are presented with cost impact. You decide whether to proceed.

Supervised
PROPOSEDSAT → 24.0°OPERATOR APPROVESevery action logged
02Supervised

Supervised operation

Weeks 7–12: commands are proposed to operators before execution. Every action is attributable and logged.

Backtest
ACTUAL HISTORYPROPOSED CONTROLREPLAYED BEFORE GOING LIVE
03Backtest

Simulation & backtest

Proposed control is replayed against the building's own history before it is trusted to run live.

Hard bounds
MAX — cannot be crossedMIN — cannot be crossed
04Hard bounds

Safety-bound validation

Hard limits on every controllable point are configured and verified with your team before go-live.

Override
AI CMDMANUALOPERATOR INSTRUCTION TAKES PRECEDENCE — ALWAYS
05Override

Human override

Operators can suspend or revert to manual BMS control at any time, without contacting us.

12 months
M1M2M3M4M5M6M7M8M9M10M11M1212 months · 68 AHUs0 INCIDENTS · 0 BOUNDS EXCEEDED
0612 months

12-month reliability

Continuous autonomous operation across 68 AHUs, zero comfort or safety incidents, no safety bound exceeded.

Safeguards

You never lose control.

Independent, layered safeguards — this is not marketing, it is how the system is built.

Immediate

Equipment firmware bounds

Physical limits enforced by the equipment itself — cannot be overridden by any software, including ours.

Immediate

Safety-bounds layer

Hard limits on every controllable parameter. The system is architecturally prevented from commanding outside them.

≤ 10s

Live state validation

Every command is checked against the building's live measured state before it is written.

≤ 10s

Kill switch

Full suspension of all system-issued commands, available to your facility team at all times.

≤ 60s

Mode flag

Revert the whole building to manual BMS control at any time, without contacting us.

Immediate

Human override

Any operator instruction takes precedence over any system command. Always. No exceptions.

Continuous

Offline-safe operation

If the connection drops, the building keeps operating safely within its bounds.

In 12+ months across 68 AHUs, no safety bound has ever been exceeded.

The frontier

The highest-stakes version: AI data centres.

Where compute is capped by power, not chips, autonomous thermal control turns freed cooling power back into usable capacity.

Roadmap — not yet in production.

Thermal → capacity
Where this goes

The physical world will run itself — one building at a time.

Built as a fleet-learning platform, designed so every building it runs is intended to make the next one smarter.

Vision. Fleet learning is on our roadmap, not a claim about today.

Get started

See a building run itself.

Tell us about your building. We'll show you the live deployment and what autonomy would look like on your equipment.

info@saskaware.com+91 8397926782

Gurugram, Haryana, India

What to expect

  • A walkthrough of the live 12-month deployment
  • What autonomous control would look like on your BMS
  • The safety model, in full
  • No obligation, no sales deck