EN
PL
Even if AI were free, waiting still costs
"Let's wait — we will soon do it ourselves with AI and it will be cheaper" — an argument that increasingly holds up investment in energy efficiency. The problem is that it loses even in the best-case scenario for those doing the waiting.
Let us assume the best case for the waiters
There is an argument running today about whether AI will be cheap. For every chart of rising spending on compute, someone will produce a chart of falling unit prices — and that argument will never end.
So I will skip it. Let us take the best possible scenario for those who wait: in two years, AI is free and better than it is today. And let us see whether that changes anything.
What will free AI do with a facility that has no data?
Picture that free, brilliant AI let loose in a facility with no metering. Connected to automation that nobody has integrated it with. With no layer able to physically switch anything — ventilation, cooling, charging. Without a single year of history on how the facility actually behaves.
What will it do? It will describe, in beautiful prose, how much money is leaking right now.
Nor will it choose a strategy for anyone. Where to start, where the lowest-hanging fruit is, what not to touch because the hall is being rebuilt next year anyway — these are decisions that require knowledge of the company's development plans, and that does not sit in any database. AI will suggest options. Responsibility for the choice stays exactly where it was.
The bottleneck is not intelligence
The "we will do it ourselves with AI" argument quietly assumes that the bottleneck is intelligence. And in energy efficiency, intelligence is paradoxically the easiest piece of the puzzle.
The bottleneck is everything around it: measurement, integration, control and experience gathered over years. That cannot be bought on subscription. It is built over time — and time runs in one direction only, much like the bills paid in the meantime.
The calculation the waiters always lose
Waiting costs — but you do not need to invoke expensive tokens to show it. It is enough to assume AI will be free, and to count what every month without data and without control costs regardless.
That calculation the waiters lose under every pricing scenario. Because when the perfect, free AI arrives, the winners will be those who already have the data, the integration and the control for it to work on.
The bottleneck in energy efficiency is not intelligence but data, integration and control built up over years — which is why even a free, brilliant AI will reward those who already have them, not those who waited for it.
Percee builds exactly the foundation without which no AI will work: metering, integration with existing automation from any manufacturer, a layer of real control, and a continuous history of data. As a finished EMOS-class system it starts saving today — deployed in stages, from non-invasive monitoring to full control — so the company does not wait, it shifts consumption, closes the loop and collects the data on which AI will one day be able to work. At Solwena's customers this means a 15–40% lower energy cost and more than PLN 50 million of documented savings across more than 50 deployments.
See how Percee works →