Participants: Malte, Benedetto, Henrik
Malte:
We’re hosting Friendsgiving in my family. It’s the second time, so I guess that makes it a tradition now. And since this is our third podcast episode, maybe we’re building a tradition here as well.
Henrik:
Nice. Yeah, we’re being consistent – at least in the beginning.
Malte:
So today we’re going to talk about supermarkets. The reason is that we’re seeing all the data coming in from these really interesting supermarket chains around Europe. There’s a lot of fascinating insights and, of course, significant savings we can identify.
If we zoom out, what excites me the most about supermarkets ties directly to our mission at Ento: reducing global energy consumption. I often joke that we’re overambitious people because we’re still a relatively small team, but we’re chasing a big mission.
And supermarkets genuinely help with that. The rule of thumb is that supermarkets in any modern country consume around three to five percent of national electricity. That’s enormous. When the numbers are that big, the return on investment is big, the financial impact is big – and that’s what we want to unpack today.
Henrik:
We also see increasing regulatory pressure, ESG requirements, and of course profitability is a huge focus area for supermarkets. When margins are razor thin, even small improvements hit the bottom line in a big way.
Malte:
Maybe we should do that calculation today. If you save one million euros on the bottom line, how much extra revenue do you need to generate to match that? How many bananas does a supermarket need to sell to earn one million on the bottom line?
Henrik:
The banana factor.
Malte:
Exactly. The energy–banana factor. Let’s trademark that.
Malte:
Another interesting point: I heard a CEO of a pension fund recently say that energy optimisation directly drives asset value. If you cut one million euros in operating expenses, business valuation can increase by a factor of twenty. Which is wild. You’re talking millions of euros or pounds in value creation just from energy savings.
Henrik:
So let’s dive in. We have a few thousand supermarkets on Ento now. Benedetto, you’ve looked at the data – what stands out this year?
Benedetto:
From a technical standpoint, supermarkets are very large consumers. In some ways, they’re comparable to data centers in terms of national consumption share. And the great thing is: we can have a huge impact using software alone.
With data centers you can optimise cooling, but you can’t stop them from computing. With large industrial plants, savings often require expensive mechanical upgrades. But with supermarkets, we can come in with software, check what’s not operating correctly, and point the energy manager to the 50 stores that will save them a million euros combined.
A metric we use is the ratio between peak daily consumption and minimum night consumption. This tells us how well systems are turned down at night. If base load is too high, something is fundamentally wrong.
Henrik:
I pulled up a slide showing this for 2,000 supermarkets. The right tail – the ones with extremely high base load – is where the gold is. That’s where big inefficiencies hide.
Benedetto:
Exactly. Those are the stores where we often see BMS overrides, ventilation systems stuck on manual, chillers running at the wrong temperature, alarms disconnected – all sorts of issues.
And all of this is detected from the main meter plus weather data. No submeters required.
Henrik:
And just to be clear for listeners: the “peak vs base load” metric uses modelled values that already remove weather effects.
Malte:
So with all that potential, it must be easy to convince supermarkets to work with us, right? Not always. The biggest misconception is submeters.
If you’ve invested millions in submeters over the years, and then we come in and say “we can get huge value from your main meter alone”, it’s natural to be skeptical.
But thanks to clients who have already saved millions, we can now show it empirically.
Another misconception is that because they have BMS systems and submeters, everything is already under control. But anyone who seriously works with energy knows that things are rarely under control.
Henrik:
A counter-argument we often hear is: “Maybe you can detect something is wrong, but how do I know where to send a technician in a huge hypermarket?”
Malte:
Supermarkets are complex – chillers, ventilation, process cooling, HVAC, bakery equipment, etc. But they’re also highly repeatable. If a chain has 500 stores, they share very similar systems.
Machine learning is perfect for this:
• First you model each building individually
• Then you compare across the portfolio
• Then you identify anomalies
Even with only a main meter, we can detect cooling setpoints for each store. That’s extremely advanced, but also extremely useful.
Henrik:
Here’s the cooling profile example we used last time. Just from main meter + weather data, we can pinpoint cooling setpoints on a massive hypermarket.
Henrik:
One question we often get is: “What can you actually find? How do I know the root cause?”
Malte:
Let me answer from the customer journey perspective, then Benedetto can go technical.
The first time a supermarket chain sees their entire portfolio analysed, the right mindset is to temporarily forget what you think you know.
If some stores use half as much energy at night as others, the question is simple:
What’s on at night in Store A that’s not on in Store B?
In 95 percent of cases, the answer is HVAC – in all its forms.
Benedetto, can you explain how the root cause analysis works?
Benedetto:
Yes. Over the past year we’ve built a root cause analysis system using both machine learning and generative AI.
The wild part is that we can deliver very accurate explanations even when we’ve never visited the building, and have only the main meter and the address.
Sometimes we also extract data from engineering reports – PDFs from physical inspections – describing equipment like heat pumps, ventilation units, refrigeration systems. Feeding that into our model gives us extremely granular root causes.
All without installing a single submeter.
Malte:
And it’s important to say: Ento is part of the total solution. Motivated energy teams, technicians, and existing systems are the ones who actually fix things. We simply give a helicopter view and highlight the biggest problems.
Across markets – UK, France, Denmark – it’s always the same pattern: overriding ventilation or HVAC systems, construction-related waste, or recommissioning issues after retrofits.
These three categories explain most of the big savings.
Henrik:
I’ve heard you mention alarms as well. These can be surprisingly important.
Benedetto:
Yes. BMS systems are often tied to alarm systems. When someone disconnects or overrides an alarm during maintenance, ventilation schedules or refrigeration controls get stuck on manual and never return to normal.
When that happens in a 4,000 m² hypermarket, the energy waste is massive.
Malte:
And this is why supermarkets are such a great fit: they have energy teams. In public sector buildings, we often hear: “These savings are huge, but we don’t have the resources to act on them.” In supermarkets, they do.
When you combine an energy team with Ento’s “AI superpowers”, the results are incredible.
Henrik:
During the energy crisis, electricity prices skyrocketed. For supermarkets running on razor-thin margins, this was a top-management issue. CFOs suddenly cared intensely about energy.
Now in the UK, even though spot prices are down, non-commodity charges are rising due to grid congestion. You can’t fight that except by reducing consumption.
This keeps energy efficiency relevant even after the crisis.
Malte:
Exactly. And honestly, I don’t know any other department where a handful of people can create millions of euros on the bottom line in such a low-margin industry.
Henrik:
Let’s address the elephant in the room. If these systems don’t fix things automatically, why not go full automation? Why not let AI control everything?
Malte:
Because supermarkets are mission-critical environments. Food safety is non-negotiable. If a control system makes a mistake and freezer temperatures drift, that’s not acceptable.
Automation platforms exist, and some have tried integrating BMS control. But most large chains are hesitant – rightly so. They stick to advisory systems rather than full AI control.
Henrik:
And to automate, you’d have to understand everything: every sensor, every system, every override, every piece of equipment. The cost and risk are enormous.
For the 1–2 times a year where something big happens, it’s perfectly fine to rely on the expertise of your technicians and energy managers.
Malte:
Exactly. And many supermarket groups approached control projects before. A lot failed. Ento is currently deployed in hundreds of stores where a previous control project had been started… and abandoned.
Henrik:
Which brings us to the bigger picture: energy waste is still the biggest opportunity. Not exotic algorithms or predictive control. Just fixing the big, stupid things that waste energy today.
Benedetto:
And that ties back to the “energy waterfall” concept:
• First eliminate waste
• Then optimise
• Then consider advanced load shifting
• Then go into distributed generation and full flexibility
Most of the world is still on step one.
Malte:
November is always busy for us. Energy teams want everything ready by January 1. And I’m heading to EMEX this week to talk to our UK clients. The narrative is always the same: everyone wants to talk net zero, but the reality is that 99 percent of the market isn’t even close.
Henrik:
We’ll end on a positive note: Benedetto has more time for coding now that event season is over.
Benedetto:
Yes, I blocked time to give the measurement and verification tool some steroids. Maybe a teaser for next time.
Henrik:
And I’m heading to Slush. Still figuring out how to hack the sauna game in Finland, but I hear it’s quite the experience.
Alright – see you next time.
Malte:
Take care.
Benedetto:
See you.
Henrik:
Bye.
.jpg)
