Why foresight is the strongest asset for the energy transition
Why we invested in Distill
Some things in our world are easy to predict: tomorrow’s rain, the working-age population in ten years, or the next solar eclipse (August 12, if you’re wondering). Energy, on the other hand, is notoriously hard to foresee, and that’s a bigger problem than most of us realise.
Electricity runs the world. So how do we not run out?
Electricity powers most of our modern world. And we need more and more of it every year. Our population is growing, and we want to keep our houses warm and the lights on. We’re further electrifying our mobility, our manufacturing, and our lives. On top of that we’re building more data centers to power advanced computing. Through 2030, global power demand is set to grow by more than 3.5% each year (IEA). In major data center hubs like Texas, the problem is more acute, with demand being expected to quadruple by 2032 (Bloomberg).
Luckily, the options to generate that needed electricity have expanded massively. Solar power, wind farms, battery plants, and other renewable generation sources have entered the mix to provide more affordable and sustainable alternatives to keep powering our lives. While the potential upside is immense, the initial investment is large. Utilities spend up to billions to get a plant up and running.
The cost of wrong predictions
Where we place a plant matters more than many people think. The physics of the grid, the weather, the supply and demand, and several other factors determine whether a plant will be a good investment. It matters so much that almost 2,000 power projects had to be canceled in 2025 in the US (Cleanview). That is almost a quarter of all the US’s current electricity generation capacity taken off the grid again. Unfortunately, this is especially true for renewable energy projects. For a lot of them, the technology was perfectly viable, but it was simply in the wrong place at the wrong time.
You might have been able to get away with placing a new power plant anywhere on the map in the past, but not anymore. While energy markets used to be relatively stable for the last 25 years, now load growth has jumped. So has the build-out of solar, wind and battery plants, which are far more intermittent than traditional fossil-fuel energy sources. All in all, the grid has become a lot more complex and changes far faster than it used to.
Decisions that don’t take these dynamics into account are often not only a financial disaster but also a waste of resources and a societal burden. Canceling projects also means losing capacity, which makes it harder to keep up with climbing demand, something all of us feel when the electricity bill arrives.
To handle that risk, energy companies usually hire one or two consultants to provide a simulation of the most likely scenario. In practice, that one forecast works a lot like a crystal ball. It’s confident, but vague, and easy to get wrong. The banks get their necessary numbers and a rough idea of what might (or might not) happen, but with this many plants having to close, the setup clearly isn’t working all that well.
The moment the future became computable
In the past, it hasn’t really been technologically feasible to do much better than that. But recently two things have changed simultaneously: the grid has gotten far more volatile, increasing the need for better prediction, and computing power has become strong enough to actually deliver it.
David, one of Distill’s founders, had spent the last 15 years close to this problem, trading power on financial markets. When he was trading for the next day, his team was running thousands of algorithmic simulations to get the best risk-adjusted return for the million dollars they’d bet on the energy market. Eventually, he wanted to use that technology for something more fulfilling: moving away from simulating market returns and toward simulating the success of physical assets. If his team could routinely run these simulations to reduce their trading risk, why would energy companies not do the same for their billion-dollar power-plant investment?
Turns out those teams don’t have a large expert technical team to do algorithmic modeling in-house, and neither do they have the means to employ a small army of consultants to create thousands of curves for them. So David and Matt, his colleague and technical lead at the trading firm, started to think: Could we take what we’re doing now and make it viable to model a power plant for the next 10 years?
From making better bets to building a better grid
By the time Sonya joined, after eighteen years building power-market tools of her own, the three of them had felt, first-hand, what it costs to be wrong about energy. They built Distill to be the tool they wished they had at the desk to derisk bets on the energy market. The platform allows their users to run grid simulations in the cloud to test various scenarios, whether they’re looking at the grid tomorrow or 10 years from now. Instead of building a whole system themselves, they simply click through the platform or call an API to explore all the relevant possibilities, while live data feeds into thousands of simultaneously running simulations in the background.
Distill trades the crystal ball for a powerful simulation engine. Instead of one best-guess forecast from a consultant, Distill hands them every plausible future at once: a load spike tomorrow, a windy year ahead, or a new data center three years down the line. The power-flow predictions capture all possible variability of the grid with factors like demand uncertainty, solar availability or unexpected transmission outages represented in the curves (check out Distill’s recent blog post to see an example of what these simulations can uncover).
The models are inherently dynamic and deeply complex, just like the grid. To get the most accurate picture, Distill combines both the physics of the grid and the market layer. Think of it like someone trying to solve the physical problem of how much power can flow through the grid at a given time and place, while an expert market analyst weighs in alongside them. Distill keeps the high fidelity resolution the problem demands but simplifies that complexity for its users with an intuitive user interface, and a programmatic API.
Simulations with impact
Today, Distill is mostly used by traders, private equity firms, and infrastructure investment companies. The platform gives them the foresight they need to make big investment decisions with a more realistic outlook on the electricity grid. As the new tech gains credibility, the focus will shift from placing financial assets well to placing physical assets, like power plants and data centers, in the right spot on the grid.
Regardless of the stakeholder, there’s a benefit of better understanding risk here: More predictability means better returns for banks on their energy assets, and more money flowing to the renewable plants we badly need. The more money available for those plants, the faster they get built. And with more foresight, utilities place those assets more intelligently, so fewer plants are put up only to be shut down again. Distill helps set off this positive feedback loop, getting us a step closer to more efficient energy production.
Renewable energy will only work if it pays off. To have it pay off, we need tech like Distill’s. The team’s ambition is to ensure we get all energy assets in the ground in the right places at the right time. We are incredibly proud of the team around David, Matt and Sonya, and are excited to see them make that happen, redefining how electricity powers the world.