Ask about combined cooling, heat and power at an AI infrastructure show and you'll hear a big number: trigeneration runs at around 87% efficiency. Among the material from Yotta 2026 was a set of slides from GenTec CHP that took that number apart, which is unusual coming from a company that sells the systems.
Where 87% comes from
A trigeneration plant runs an engine that always makes electricity, and puts its waste heat to work: heating in winter, and in summer, cooling through an absorption chiller. Add up the electricity, the heat and the cooling, divide by the fuel, and you get the headline figure.
The problem is the addition. That math counts a kilowatt-hour of cooling made from waste heat as worth the same as a kilowatt-hour of electricity. It isn't. Electricity can do anything; low-grade heat and the cooling it drives can do much less.
Counting it honestly
Converted to an electrical equivalent, the slides argued, the same system lands at about 52%. Trigeneration then adds roughly 5 to 10 effective points over plain cogeneration, not the 30 or more the headline implies.
That doesn't make it a bad idea. The same slides pointed to retrofit case studies with roughly 20% lower primary fuel use, a real saving. It just makes it a different idea from the one the brochure sells: an efficiency gain worth having, not a near-doubling.
What we took from it
For a data center weighing on-site power and heat reuse, the honest number is the one that survives contact with the utility bill. That means measuring what the plant actually delivers, electricity, heat and cooling, hour by hour, against the fuel it burns, and valuing each for what it is. It's also a reminder to ask the same question of every efficiency figure on a show floor: what was added to what?
