Same hardware, different P&L: why two identical batteries end up 30% apart
Twin batteries from the same manufacturer, with the same power rating, the same connection point and the same warranty, can finish the year up to 30% apart in revenue depending on who operates them. The storage panel at RENMAD UsefulAI 2026 put a number on that gap and showed where it is won — in the small print of the contract you sign with your operator.

Picture two batteries, twins from the same manufacturer, with the same duration, the same connection point and the same warranty curve from the same supplier. Put them side by side and no engineer could tell them apart. By the end of the financial year, one will have earned up to 30% more than the other. The only difference is who operates them.
At the session in Madrid on 2 June 2026, Cristina Corchero, technical director and founder of Bamboo Energy, asked whether the same hardware can perform completely differently depending on the quality of the software. Juan Galiardo, Entrix's Regional Manager for Iberia, said yes, on one condition: you have to compare "apples with apples", because each asset's constraints can have an enormous impact. Compare them that way and "we're not talking about a small difference of 1 or 2 percentage points, we're talking about cases where there can be deltas of 20 or 30%". He says he sees that delta in Germany, where Entrix has been operating batteries for five years.
The calculation that has to be redone every 15 minutes
A battery earns in several markets at once: day-ahead, continuous intraday, secondary reserve (aFRR) capacity and activation, and voltage control, each with its own window and its own price. A megawatt you sell in one is a megawatt you can no longer sell in another. The calculation has to be redone four times an hour for the entire life of the asset, taking account of state of charge, degradation and liquidity. Liquidity as thin as Spain's forces you to split volume into smaller orders so that they get matched. Run a portfolio and you have to optimise each asset, then the portfolio as a whole. "In other words, game theory," Galiardo summed up. He gave his conclusion in the first person. "I think asking a trading team to execute that in real time, without error, is basically impossible."
Asked which of price forecasting, optimisation and execution he would put the budget into over the next two years, Galiardo gave an answer the moderator immediately labelled disruptive, and qualified it as holding for now: "if I had to put weight on any of them at this stage, I'd put it much more on optimisation and execution than on price prediction". Spain, he said, has enormous experience in price forecasting, and maybe there is less to add there. What pays is coordination, because "you could have an incredible price forecast, but if the execution then isn't done in an automated way and an alert goes off for someone and somebody has to be watching, hey, but does this order go in or not, it's already too late".
The gap widens in the lean years
This is where the 30% stops being an operational curiosity and becomes an investment problem. Galiardo framed it as an expectation. When the market is handing out a lot of money, maybe the delta in percentage terms is lower, because with margin everywhere it is easier to get things right. When revenues tighten, he thinks it will be bigger. And the lean years are exactly the ones your financial model cannot absorb.
Public revenue indices track the swings month by month. In Great Britain, the benchmark battery went from about £41,000/MW/year in February 2026 to about £70,000 in March, up 69% in a month. Behind the jump, day-ahead spreads widened from £40 to £89/MWh and the contribution from the Balancing Mechanism doubled. If revenues swing like that from one month to the next, what decides your year is being in the right market in the right quarter-hour.
In Spain the problem comes from the other direction. By 29 September 2026, the Spanish day-ahead market had clocked up 906 hours with the price at or below zero, concentrated in the middle of the day, when the sun is up. That is our own count, from the daily files of OMIE, the Iberian power market operator. UNEF, the Spanish solar association, put this summer's evening wholesale price at €250/MWh. Public Iberian arbitrage benchmarks put 2025 at about €130,000/MW, though those are theoretical capture figures, not revenue actually earned. On that basis, a delta of twenty or thirty points is worth tens of thousands of euros per MW per year. "Look, I'm a manufacturer and it's in my interest to sell batteries, but it's clear that in the future in the end the real value you're going to get out of it is in trading and the battery is becoming commoditised and that's true," said Juan Álvarez, Senior Solutions Manager at Envision. There is a limit to that. The BMS, the system that protects the cells, has the final say, and it can override a market order. Operations decide within those limits.
That 30% can be written into a contract
The same gap that scares lenders off today is something you can negotiate in your contracts. On 21 September 2026, UNEF said that by its reckoning the list of lenders financing storage stood "at zero", noting that until recently there had been one. It proposed that the ICO, Spain's state financial agency, open a line of about €315 million a year. Operating skill is something you specify and pay for.
Start with the warranty, where you tie your own hands. Galiardo insisted that "it's so important that the warranty parameters aren't defined as daily cycling, but as average annual cycling, because that's what allows you to cycle". With a daily cap you cover the same mileage over the year, but you cannot cycle more in summer and less in winter. The scheme Álvarez described starts from one cycle a day, allows a flexible warranty of one to two cycles on average and tolerates three, provided it is not continuous. It also adjusts state of health according to throughput, the megawatt hours you have actually taken out of the battery. That part involves no artificial intelligence, Álvarez admitted. It is modelling that ends up in Excel.
Next comes the decision function, the formula the optimiser uses to choose between opportunities. José García Franquelo, of Isotrol, explained how they build degradation into it in the unit an investment committee understands, because "that degradation is expressed in euros". Corchero noted that not every European operator does this today. Isotrol's clients, García Franquelo said, ask for discipline except when the price jumps far enough to justify breaking it. They tell the company "set the fixed rules for me, give me what the manufacturer says […] preserve its useful life and of course preserve the warranties. But if there's a sprint of 150 euros, go all out." What separates one optimiser from another, he said, citing the firm that publishes the indices, is catching the opportunities that come up rarely and pay a lot.
What to demand of your operator
The sector already knows how to turn operating parameters into bankability: stretching a guaranteed curve from 20 to 25 years is financial arithmetic. "In the financial model going from you putting in 20 years to putting in 25, the IRR goes up and that makes it more bankable," Álvarez explained. The operating contract is the next one that has to do that job. Negotiate cycling as an annual average and insist that degradation goes into the decision function valued in euros. Benchmark your operator against an independent index, and check whether it catches the rare spikes as well as getting the average right. And ask whether state-of-charge estimation comes as standard or has to be bought separately, because state of charge is not measured, it is estimated, and the error accumulates. On that point, Galiardo said that if it moves the needle by two or three points, "for a 100-megawatt battery that's a lot of money".
Accept, too, the autonomy this demands. Asked to score themselves from 0 to 3, none went below 2. Galiardo said "we have to aim for three, but with human supervision", always with a team watching 24/7 that can take control. And he put a price on the validate button. "But we do believe that the need for order-by-order validation takes a great deal of agility away and therefore leaves money on the table." García Franquelo put himself "at two and a half".
When Corchero asked them to split the value of a battery in 2030 between the electrochemistry and the intelligence that operates it, none of them sided with the hardware. The answers were 90/10, 85/15 and 80/20, with operations always taking the larger share, and the 85/15 was Galiardo's. Then she gave them a sentence to finish: the decisive competitive advantage in these batteries will not be having more megawatts, but… "well-used megawatt hours".
The hardware arrives on a lorry, delivery note and all. That 30% arrives in the operating contract, and that is the one you sign yourself.
Storage panel at RENMAD UsefulAI 2026, Madrid, 2 June 2026. Juan Galiardo, Regional Manager for Iberia at Entrix; Juan Álvarez, Senior Solutions Manager at Envision; José García Franquelo, of Isotrol. Moderated by Cristina Corchero, technical director and founder of Bamboo Energy.
You reached this analysis because ATA Insights puts independent energy-transition intelligence in front of 86,000+ professionals across the sector. That's exactly what we do for funded projects, events and companies — turn your work into reach.
See how we disseminate work like this →Where this analysis came from
This piece draws on RENMAD UsefulAI 2026, 2 June 2026, Madrid.
The energy transition, in your inbox
Join 86,000+ professionals reading our independent briefings across five sectors.
