Energy Efficiency or Uptime First?

When the power fails in a grocery store, the freezers stop, the food is wasted and the tills go down. A founder whose customers include grocery stores said that risk has become a much bigger worry this year, and it is what gets the company through the door. Efficiency products have been on offer to the same shops for years. Several practitioners at Jan Palasinski and Adele Unneberg's Ripple on industrial buyers told a similar story. Manufacturers buy uptime and lower costs, and the carbon cut comes along with them.

Sold on uptime

Jan, a partner at Future Energy Ventures, said the software start-ups that have really made it with manufacturers can be counted on one hand. Tractian and Augury in the US, and Sensorfact in Europe, which ABB acquired, all sold the same thing. They attached sensors to machines, gathered data and promised higher uptime. An hour of downtime costs €200,000 to €1m, depending on whether the plant is in process industry or discrete manufacturing, and in process industry the emissions carry on while output is zero.

Selling software used to mean long projects with system integrators that cost millions. Jan said the time to value, once measured in months, is now two to three weeks, thanks to agentic AI, semantic data layers and digital twins. Now that everyone uses AI at home, manufacturers no longer think it is a scam, and software can fill the gap as the experts who run their plants retire. Efficiency usually comes from the same data, so the customer buys uptime and gets the energy savings with it. Only in process industries such as steel, Jan said, is energy central enough to the P&L to be the top priority.

it's not a hardware versus software, it's a hardware and software.

The first customer

Adele, of Sandwater, looked at the same problem from the hardware side. Capex slows procurement, but the return is easier to calculate, so much gas displaced by so much variable electricity. Modular systems let a plant electrify step by step and keep its gas boiler as a backup, which eases the fear of downtime. What sells the next system is a lighthouse customer whose kit has run for thousands of hours without fault. In Adele's experience, that first customer has to want to decarbonise and be willing to put a value on greenness.

In Germany, one participant said, grants can bring a first customer's capex close to zero, and a design with redundancy, which cannot halt the production line if something goes wrong, makes it easy to say yes. One of the hosts concluded that the first sale rarely rests on economics alone. Either the customer wants to decarbonise or a government pays it to. Another participant disagreed. Decarbonisation is not a factor right now, they said. Manufacturers are under intense competitive pressure and buy for the economics and for less exposure to volatile fossil fuel prices. Someone asked whether it had ever been a factor.

An investor said large manufacturers, even those with ten factories, often have no clear view of their energy use. A sale then involves the plant manager, who owns the local economics, a global facility management team and the sustainability team, and one of the hosts added IT security. A founder has to work out who really decides, the investor said, and remember that "Cost is king." An investor from an energy company advised start-ups to join a bundled offer, because a utility with long-standing customers has the credibility to bring a start-up in.

Hardware or software first

The hosts took opposite sides on what should lead. Sandwater treats proprietary technology as the main moat, Adele said, which often leads to hardware. It is harder to sell, but once a large machine is installed, it is physically sticky. Jan agreed that hardware drives the cuts in emissions and cost, but saw software as the enabler. A digital twin shows a plant what electrification makes possible, and software then decides when to store heat and how to trade flexibility. Once a site is electric, 90% of the job is done, because the grid keeps getting cleaner. If software shows a four-year payback from trading a flexible asset, the hardware decision becomes a no-brainer.

An investor saw it the other way round. You cannot electrify a gas furnace with software. The electric boiler comes first, then the software for trading, batteries and thermal storage. A hardware maker that offers software to steer its own asset is in a stronger position than an asset-light software firm asking to control equipment someone else installed, they said, especially now that AI makes software layers easier to build. A founder working on edge autonomy, where compute and power are scarce, said the two have to be designed together, and cited NVIDIA's Jetson as fine hardware held back by weak software.

Adele closed by saying everyone could agree that "it's not a hardware versus software, it's a hardware and software." Jan had already offered the commercial version. If software shows a factory what electrification will pay, the factory buys the hardware, and Jan joked that the software investments would earn Adele a huge carry. In both versions, the carbon comes as a side effect.

This Ripple was hosted by Jan Palasinski (Future Energy Ventures) and Adele Unneberg (Sandwater) at The Drop 2026 on 16 September.