Chiller-plant tuning for hospitals and malls, paid from measured savings
TL;DR Chillsteady fits meters and a small gateway to the central chilled-water plant of a hospital, mall or office park, and then adjusts set points, pump speeds and chiller sequencing hour by hour. The owner pays a small fixed fee plus a share of the measured power saved.
Cooling is the fastest-growing slice of India's energy use. The India Cooling Action Plan expects building space cooling to grow almost elevenfold by 2037-38 and to make up 74% of the country's cooling demand[1]. Worldwide, air conditioners and fans already use nearly a fifth of the electricity consumed in buildings[2].
In a large Indian building the chiller plant is usually the single biggest load, yet many plants run on fixed set points chosen at commissioning and never touched again. Night shifts, part-load hours and monsoon humidity all get the same settings, which is a bit like driving in one gear all day.
Chillsteady starts with a two-week survey - power meters on chillers, pumps and cooling towers, plus temperature and flow sensors where the building management system doesn't have them. Models learn how that plant behaves at each load and weather condition. The controller then sends small, bounded changes through the existing system, like raising chilled-water temperature at part load, running pumps faster or slower and picking the most efficient chiller combination. Facility staff can override any change.
Each site's raw data stays on its own gateway and only model updates travel, so lessons from fifty plants improve the fifty-first without owners sharing their data. That was the seed's federated learning pitch.
I've reframed the seed though. Owners don't buy a learning method, they buy a lower power bill they can verify. So the product is sold on measured kilowatt-hours saved against a baseline both sides sign.
The facility head uses it but the CFO or owner signs. Hospitals run cooling around the clock, malls carry large common-area loads and IT parks pass costs to tenants. A fixed monthly fee covers hardware, and a share of verified savings pays the rest. I found no public Indian benchmark for these fees, so test the split.
Commercial buildings named as designated consumers must have an accredited auditor check their energy use within 18 months of notification and every three years after, with an energy performance index per square metre[3]. The building code has moved from ECBC 2017 to ECSBC 2024[4]. Both put energy numbers in front of owners more often.
Nobody publishes a figure for India's chiller-optimisation spend. The planning target is the clearest signal - the Cooling Action Plan aims to cut cooling energy needs by 25–40% against its reference path by 2037-38[1]. Treat that as policy intent, not revenue. Build your own count of hospitals and malls with central chillers in one city before raising money.
One AI HVAC firm, now part of a big HVAC group, sells autonomous HVAC control and a cloud building system, and one retailer grew from 49 stores to 214[5]. Energy-savings firms work on commercial building energy savings in India[6]. Large chiller makers and BMS vendors also sell optimisation as an add-on. So the gap is a brand-neutral service for mid-size plants, priced on savings, that works across mixed equipment.
| Difficulty | Moderate (48/100) |
| Starting capital | ₹1.2–2.2 Cr |
| Time to launch | 9–12 months |
The hard parts live in the plant room, where every site has different chillers, old controls and cautious engineers. Agreeing a fair baseline that adjusts for weather and occupancy decides whether you get paid at all. Field installs and remote support across cities also take steady operations work.
No sector licence is needed to tune chiller plants. A few energy and hardware rules shape the work.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't let AI run a hospital plant without limits or human override. Comfort and critical areas come before savings, and bad sensor data can mislead any model.
Want to work out the AI layer for Chillsteady? Talk to Yash →
I think it's a solid business for a founder with HVAC commissioning or energy-services experience and some controls engineering. If you're a pure software team that has never stood in a plant room then skip it. Prove one number first - verified savings above 10% of chiller-plant power at three sites for three months, with no comfort complaints.
Plan for roughly ₹1.2–2.2 Cr, much of it sensors, engineers and the cash gap while savings-based fees come in.
No sector licence. Use gateways that already carry WPC and BIS approvals, and BEE audit rules can help you sell.
Yes, anywhere with large air-conditioned buildings and high power bills. Building codes and energy audit rules vary, so check those locally.
The free Chillsteady validation kit has the line-by-line capital breakdown, the effort scoring, the energy-audit and data checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for facility heads, and 5 assumptions to test. Download the kit →
Thinking about building Chillsteady or something close to it? The make-or-break step is a savings baseline that both you and the owner trust. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.