A FEBRUARY ARTICLE IN FINANCE RUINED MY COFFEE
It was a Tuesday morning. I had coffee in one hand, my PIES 2026 slides open on the other screen, and a genuinely optimistic plan for the day.
Then the Finance article appeared.
Nataša Koražija, 24 February 2026. Headline: Ponovna menjava elektroštevcev: tehnologija ne ustreza novim potrebam napovedovanja. Translation: we are doing the whole electricity meter thing again. The ten-year replacement cycle that just ended? Yeah. About that.
ELES, quoted directly: roughly €150 million to replace all the meters. 800.000 electricity meters across the five distribution companies. Ten-year project. And, this is the part that made the coffee cold, the €150 million does not include installation or mobile communication service costs. Nice!
Three sentences that rearranged my week.
By lunch I was no longer preparing a PIES presentation about contextual metering. I was preparing a PIES presentation about contextual metering with a new chapter called "who gets paid for all of this." Because once you start counting, you cannot stop.
And I ran out of coffee. I need to stop drinking coffee anyway.
THE METERS ARE FINE. THE POWER LINE IS NOT.
Before we count the money, a quick sanity check on what is actually broken.
The meters are not broken. ELES says this out loud in the Finance article: "tehnologija števcev, ki smo jih nameščali od leta 2018, omogoča vse potrebne funkcionalnosti". The boxes on the walls are fine. The boxes on the walls measure things correctly.
What is broken is the thing that connects those boxes to the distribution company. That thing is PLC - Power Line Carrier. You send the data through the same wire that carries the electricity. Brilliant in 2016, because you did not need new infrastructure. Painful in 2026, because the low-voltage network is now full of people.
People with solar panels, heat pumps, batteries, EV chargers. And - I love that the Finance article names this specifically - people with cheap power supplies from Chinese online stores. Every one of those devices generates electrical noise. Individually within regulatory limits. Collectively, the sum of all that noise is strong enough to make PLC give up.
Today, only about 6% of Slovenian meters can send data over mobile networks. The rest still yell into a power line that has become a very busy, very noisy street. And that is why your Moj Elektro data arrives with a 24-hour delay. Not because of policy. Not because of privacy. Because the pipe is too narrow and too noisy for anything faster.
ELES said this. I am not making it up, and I am not being unfair to anyone. It is just the situation.
SO I PREPARED A PRESENTATION FOR PIES 2026
I had a plan. A good plan, I thought.
Slide 1: today we have 15-minute intervals with a 24-hour delay. Slide 5: the 30-second cloud scenario that kills transformers. Slide 6: a big number about data volumes. Slide 10: Vela architecture - CASSIOPEIA, Scorpio, LIBRA, NGSI-LD, the works. Slide 14: a pilot proposal, 120 measurement points, 90 days, reasonable budget.
Very technical. Very clever (well, I thought so). Very proud of the context broker.
Then I read the Finance article a second time, because something was bothering me and I could not say what.
It was this: the article answered "what will we build" and "when will we build it" and "why does the old thing not work anymore." It did not answer one question. The question I should have been asking for months. The question nobody was asking.
Who gets paid?
WHO ACTUALLY GETS PAID?
When you see a €150M number in the newspaper next to a ten-year project, the instinct is to treat it as one big pot that "the industry" collects.
It is not a pot but six separate dinner tables, each with a different menu and a different bill.
- The people who make the meters.
- The people who rent you the SIM cards that go into the meters.
- The people who install the meters.
- The people who make the little communication modules inside the meters.
- The people who store, process, and secure the mountain of data coming out of the meters.
- The people who build the platform layer that turns that data into something useful.
I am in the last group. I was going to my own presentation to sell table six. Which is fine. I like table six. But I wanted to know, honestly, what the other five tables were eating.
So I sat down, I opened a spreadsheet, and I started counting. Out loud. The rest of this post is the counting.
Fair warning: some tables eat a lot more than I expected.
THE METER MANUFACTURERS: A BIG ONE-SHOT
ELES's €150M is hardware only. That money lands - almost entirely - on whoever wins the national tender to supply 800.000 meters.
In Slovenia, the obvious candidate is Iskraemeco. Headquartered in Kranj, part of the Elsewedy Electric Group since 2007, the incumbent for the 2020-2022 rollout of the AM550 meters that most Slovenian households already have. They know the country, the regulator knows them, the distribution companies know them. If the next tender has "Iskraemeco" written at the top of the winning envelope, nobody will fall off their chair.
Landis+Gyr (Swiss-listed, global) is also the usual alternative in EU tenders, and there are a handful of others. The point is not who exactly wins. The point is the shape of the money.
€150M over roughly ten years. Big. Visible. Headline-worthy. Press releases, ribbon cuttings, a genuine boost to a national manufacturer. All good things.
But there is a catch, and it is the same catch every hardware manufacturer lives with: when the last meter ships, the invoice stops. The next big order is another ten or fifteen years away. You have to eat very well during the feast, because the feast ends.
This is not a criticism of meter manufacturers. It is the nature of hardware. Now hold this shape in your head - big, one-time, visible - because we are about to meet the opposite.
THE TELCOS: A SMALLER NUMBER THAT NEVER STOPS
Here is where it gets interesting.
ELES was very specific in the Finance article: the €150M does not include mobile communication service costs. That line is separate. That line is someone else's invoice.
Those 800.000 new meters need to talk over mobile networks - NB-IoT or LTE-M, the two low-power cellular flavours designed for exactly this job. Every single one of them needs a SIM card. Or an eSIM. Or an iSIM soldered directly to the board, which is the modern way. And every single one of those little pieces of silicon needs a monthly subscription.
How much does that cost? This is where I stopped guessing and started looking at public tariffs.
The European IoT connectivity market has public pricing, and it lives in a pretty narrow band. 1NCE offers €10 for ten years of coverage on a single SIM at their flat rate - works out to under €0.10 per month. Onomondo, Olivia Wireless, A1 Digital, Deutsche Telekom IoT - all price utility-grade NB-IoT service in a range of roughly €0.30 to €1.50 per SIM per month, depending on volume, data allowance, and contract length. National DSO tenders typically land in the low end of that range because the volumes are large and the traffic is tiny.
So let us do the math out loud, with your own calculator if you want.
Take the middle of the public range: €1 per SIM per month.
800.000 SIMs × €1 × 12 months = €9,6 million per year. Over ten years: €96 million.
At €0,50 per SIM per month (aggressive utility tender pricing): €48 million over ten years. At €1,50 (premium, no real tender discipline): €144 million over ten years.
I am not going to claim the exact number. I don't know the exact number. Nobody outside the tender room will know the exact number for at least two more years. But look at the shape.
Somewhere between €48 million and €144 million, in one small country, for one infrastructure project, flowing to whichever telcos win the contract. In Slovenia that means some combination of Telekom Slovenije, A1 Slovenija, and Telemach/T-2. Smart companies, all of them. Playing the long game.
And here is the part that matters. Go back and look at the meter manufacturers. €150M. Big number. Eats for maybe four years of delivery, then done.
The telco number is smaller in gross, but it does not stop. Month one of year one looks exactly like month one of year ten. And when AMI 3.0 shows up in 2036 - because it will - the telcos will still be billing while everyone else is pitching the next tender.
A smaller pie. That never stops being a pie.
The thing about recurring revenue is not the size of any individual bill. It is that you never have to send the first one again.
INSTALLERS, MODULE MAKERS, AND THE PEOPLE WITH RACKS
Three smaller tables in one chapter, because the honest answer for all three is "meaningful, but I do not have good numbers."
The installers. ELES said out loud that installation labour is excluded from the €150M. Installation is its own line. Across Europe the per-meter installation cost varies wildly - urban vs rural, ground floor vs cellar, single-phase vs three-phase, in-house crew vs outsourced contractor. The honest answer is "tens of millions of euros, somewhere." Whoever gets the installation contract will be busy for years. Unglamorous, unrecurring, but solid money for the people with ladders and trucks.
The module makers. Inside each new meter is a little cellular modem - a piece of silicon from Quectel, u-blox, Nordic Semiconductor, Fibocom, Telit, or one of their peers. Those modules are not the biggest slice of the meter BOM, but they are real. Every meter needs one. Module makers eat when meter makers eat, in a quiet supply-chain kind of way.
The data center and ICT integrators. Somebody has to store, process, and secure everything that comes out of those SIMs. Rack space, cooling, backup, NIS2-compliant cybersecurity, DSO-grade SLAs. I am going to give this table its own conversation in Part 2 of this post, because the data side has its own plot twist.
All three tables eat. I just cannot tell you exactly how much, and I am not going to invent a number.
AND THEN THERE IS US
Last table. Our table.
We build the platform layer. The NGSI-LD context broker. CASSIOPEIA that translates DLMS/COSEM into Smart Data Models. LIBRA that gates access with proper policy enforcement. The clever part that turns a firehose of meter readings into something a DSO can actually forecast with.
It is the strategic layer. It is the layer that decides whether €150M of new meters turns into flexibility markets, dynamic tariffs, and proper grid intelligence - or whether it turns into the same database we have today, just bigger.
It is also, invariably, the smallest invoice on the project.
This is not me complaining. This is the pattern. Everywhere in Europe. Hardware is a big CAPEX line. Connectivity is a big OPEX line. Software is the line that gets negotiated hardest and shipped last. I have been in this business long enough to see this show three times.
I am not bitter about our position. I am honest about it. Somebody has to be the clever small slice. We are good at that.
THE UNCOMFORTABLE LEAGUE TABLE
If I rank the six tables by what matters to a business - not just revenue, but how sticky it is, how predictable, and how quietly it grows - the picture stops being flattering to the people with the biggest headline number.
The meter manufacturers win on gross revenue. One big contract, one big press release, one genuinely important milestone for the national industry. But the money stops the day the last meter is installed, and the next comparable order is a decade away. Big, visible, finite.
The telcos win on almost everything else. Smaller gross number, but recurring month after month. Extraordinarily sticky - once 800.000 meters are provisioned on a specific carrier's APN, nobody is switching providers in year four because somebody sent a better quote. High margin, long contracts, and the quietest marketing of any table in the room. Nobody wrote the Finance article about the connectivity tender that will eventually follow this one. And nobody will.
The installers eat well during the rollout phase and then go back to their normal calendar. Solid, honest, unrecurring work. Similar shape for the module makers, who quietly ride along in the meter vendor's supply chain - a real line of business, but not a separate conversation in the news.
The data center and ICT integrators live somewhere in between. Some of their revenue is recurring (hosting, operations, SLAs), some is project-based (initial build-out, migrations, upgrades). Respectable, useful, unglamorous.
And then there is us. Smallest invoice. Highest strategic leverage. Zero silence - we talk about what we do, which is apparently a personality flaw in this industry. Software is where the intelligence lives and where the margin does not.
Finance wrote 1.500 words about the meter swap. Well-reported, accurate, necessary journalism. But the story underneath the story is that the quietest player in the whole chain has the best business model.
Nobody is ashamed of this. The telcos did not invent recurring revenue, they just understood it earlier than anyone else in this industry. The meter manufacturers did not choose hardware cycles - physics did. And we, the software people, chose software with our eyes open.
The shape of the value chain is not a villain. It is just math.
THE KISS CONCLUSION: INVEST 20%, GAIN 80%, BUT MAYBE I SHOULD START A TELCO?
If you have read any of our KISS posts before, you know this is where the Pareto principle usually shows up. Invest 20%, gain 80%. It has been our closing beat since this series started.
But this time, after writing all of the above, I am looking at my own slide deck with a new question. And I want to share the question honestly, because I think it is funny and also a little bit alarming.
If the pattern is:
- Meter makers eat big, once
- Telcos eat forever
- Installers, modules, and data centers eat somewhere in between
- Software eats last
Then maybe the real KISS takeaway is not "invest 20%, gain 80%." Maybe it is:
Invest 20% in software. Invest the other 80% in a side hustle as an eSIM provider. And you will be finally financial strong enough to compete in a software domain.
I am not joking. I am slightly joking. Well, maybe 30% joking.
Imagine the LinkedIn post. "Excited to announce that SenLab is pivoting from context brokers to IoT connectivity. Our new eSIM product, powered by standards we do not understand yet, is now shipping to utilities across the EU. Subscribe for €0.87/month, forever."
I could call it VelaSIM. I could put an NGSI-LD sticker on the packaging. I could launch a co-branded eSIM with somebody who already has an MVNO license, take a small cut, and quietly send every customer an additional tiny invoice every month until the end of time.
And the honest truth - the thing I am choosing to find funny rather than depressing - is that this would probably make me more money than the software.
I am not going to do it. I like building context brokers, data hubs and similar stuff. I like the part of this IoT industry that turns data into decisions. I like the technical problem more than I like the billing problem.
But if somebody at Telekom Slovenije, A1, or T-2 is reading this - and statistically, somebody is - please consider this my formal notice that I am available for coffee, and that I have some ideas.
Mostly about the software. Mostly.
ONE MORE THING - THIS IS NOT JUST A SLOVENIAN STORY
Everything I wrote above is specific to Slovenia, because Slovenia is where I live, where I read Finance at breakfast, and where the numbers are small enough to fit into a single article without losing the reader somewhere between the fifth zero and the sixth. Eight hundred thousand meters, one hundred and fifty million euros, five distribution companies, three (ish) telcos, one Tuesday morning article that ruined my coffee.
It is a small country, which means the whole situation can be described in one long paragraph without anyone needing to squint.
But if you swap the country name, replace the company names, adjust the zeros at the end, and run the same arithmetic, you end up looking at almost exactly the same picture in every other EU member state. Germany is working through it on a much larger scale and with its own cultural approach to privacy and certification, France has already been through a full first cycle with Linky and is now living with the consequences and opportunities of that deployment, the Netherlands and the Nordics finished their rollouts years ago and are quietly budgeting for the next round, Spain and Italy are somewhere in between, and the countries that started later are moving fast to catch up.
The technology stacks, regulatory conversations, politics differ, and the timelines are offset by a few years here and a decade there, but the shape of the value chain is remarkably consistent from one border to the next.
And this is not only an EU story. Every country that runs an electricity grid is somewhere in its own version of this cycle, whether that is a first national rollout, a second-generation replacement, or an early conversation about what comes after the meters people already have. The six tables I described are always in the room. The meter manufacturers carry the headline, the telcos carry the monthly invoice, the installers carry the ladders, the module makers carry the silicon, the data center people carry the racks, and the software people carry the clever part that decides whether all of this becomes useful or just more expensive.
The names on the nameplates change, the flags change, the acronyms change, but the structure holds.
So before you close this tab, do the small exercise for your own country. Take the number of electricity meters where you live, multiply by something reasonable in the public IoT tariff range, multiply by a hundred and twenty months, and see where you land.
You do not need a spreadsheet. You need a napkin. And then tell me, honestly, whether the number you arrive at surprises you as much as mine surprised me.
DISAGREE, COMMENT, OR WISH TO KNOW MORE?
A BUS COMPANY SAYS THEY'RE ON TIME. WE GAVE THE CITY A WAY TO VERIFY.
