Software eating the grid, finally
In short
To sell grid software, make it something the operator can capitalise
David Delfassy and Moritz Müller asked where value will accrue as the grid goes digital, and people who sell to grid operators answered with how those buyers make money.
One grid-planning product was sold to about 90% of Finnish utilities and made their planning ten times faster. The utilities still said they had no time to focus on it. The participant who told the story at the grid software Ripple had worked with distribution system operators (DSOs) for more than 15 years. The hosts wanted to know where value will accrue as the grid goes digital. Most of the session went on an earlier question, how a grid operator decides to buy anything at all.
Ninety-nine problems
Moritz Müller of SET Ventures had framed the session around the grid operator as a buyer. The veteran's rule was that "they have 99 problems and yours, the one you're solving, is probably not one of them." Grids in Europe work too well to force change. DSOs like innovation, and their R&D teams love pilots, but a pilot can stay R&D for good. Replacement cycles run five to ten years, so a vendor may sell for three to five years before a deal lands. A product has to reach the chief executive's top five priorities, and nobody builds a business on DSO pilots without a lot of money.
A founder building energy-management software asked whether DSOs were now more open. Much of their work is document processing done by hand, they said, and some aggregators cannot join flexibility markets because the application queue is full. The veteran called DSOs anti-innovation organisations, for good reason, though changing a little. Once won, they are the best customers to have. A participant who once ran grid operations at a utility remembered the motto there, in Germany at least, as keeping the lights on and beating the regulator.
Where the margin sits
The most practical advice came from a founder whose software finds reliability and resilience improvements for US utilities. They sold to one of the largest utilities first, because the others watch the largest and follow. Then they aimed at capital spending. Utilities earn a margin on their capital budget, while operating money means scrambling each year for maintenance funds. If the analysis is tied to upgrades the utility is making, the utility can capitalise it. A reusable data asset can partly count too, depending on accounting standards. Failing that, add hardware, which is easier to capitalise.
Another participant explained how operating savings can backfire. A DSO makes its money on a regulated basis, and in some countries the regulator forces it to cut its operating budget every year. A tool that lowers operating cost can hurt the buyer. Moving the reliability indices SAIDI and SAIFI helps. So does selling through the suppliers DSOs already use, since a package bought from ABB or a similar vendor can carry software into the capital spend. The former grid-operations manager had sat with the hardware camp for the same reason. Hardware, they said, is good for a grid operator's balance sheet.
the only way we find out that something's broken is because we either see a lot of smoke or a lot of angry people are calling us.
Pilots that pay
Investors had mixed luck with pilots. One fund backed utility-software companies in 2016–2018 that ran interesting pilots, and all of them died. It now waits for ongoing commercial sales, which, with some of the heat gone out of climate tech, no longer means paying 2022 valuations.
Another investor had a happier story. In a recent grid-monitoring deal, pilots paid 10 to 15 times the going rate and brought in more than a million in commitments. The fear was the tender. DSOs are semi-public, so large contracts go to tender, and the fund had invested ahead of a big one. It took comfort from a list of other operators tendering the next year, and from the company's habit of inviting peer operators to watch the pilot live.
Getting in and staying in
On the hosts' opening question, the DSO veteran said value will always accrue to the DSO. Only behind-the-meter assets change that, because less grid means less capital earning a regulated return. Others doubted that software could do much alone. One participant recalled a DSO telling them that "the only way we find out that something's broken is because we either see a lot of smoke or a lot of angry people are calling us." A transformer built in the 1960s does not know what software is.
David Delfassy of TDK Ventures raised private grids, designed around software from the start, which sell power to large industrial customers while still trading with the public network. One participant objected. Where they come from, $100bn is tied up in a grid running at 50% of capacity. The obvious target is the idle half, not a parallel network.
Defensibility came back to getting in. With the cost of writing code heading towards zero, one participant said, hardware controls the connection points and the proprietary data that AI needs. An investor described a metering company whose meters last 20 years on a first contract of eight. Switching means ripping out meters that would work for another 12. Someone else said code they wrote 20 years ago still runs at some utilities. Two software founders answered that their moat is simplicity and accumulated context. Once a platform holds a user's history, leaving it means starting again.
David closed still undecided on where to invest. They had made one private-grid investment, which the objection had just cast in doubt, and said they were looking for more.
This Ripple was hosted by David Delfassy (TDK Ventures) and Moritz Müller (SET Ventures) at The Drop 2026 on 16 September.