Welcome to the second in our series of discussions led by Gorilla CEO Ruben Van den Bossche and VP of Product Joris Van Genechten. Last time, Ruben and Joris came together to give their thoughts on the big questions facing energy retailers based on their own experiences talking with retailers around the world. In this part, the pair are looking at the evolutions taking place in electricity grids, as growing demand drives innovative new solutions and distributed energy resources continue to proliferate. How will an increasingly complex grid impact the challenges that retailers are already facing? What is needed for accurate forecasting in this new world?

A small step or a giant leap for the grid?

Ruben: One interesting observation from the energy industry very recently is that large tech companies like Microsoft and Google started to invest in building their own nuclear power stations. What are your thoughts on this?

Joris: I think in those companies that have very energy intensive processes, energy is becoming more of a strategic asset. So being able to create a reliable, stable, predictable energy source could actually be a real benefit both in going green, and also in the cost of managing it.

Ruben: I think it is a big leap, right? People were surprised, the people we speak to in the energy retail space. Going into nuclear is a huge leap from installing some windmills on your data centre or going for a long-term Power Purchase Agreement (PPA) with a solar farm. 

I think what's interesting for me is that it is happening all over. It's that scale, they're at the scale of consuming energy that no one else is close to. AI is using energy at an unseen scale at a time when energy is still very scarce and when we are going into a transition. 

But we're seeing this at a smaller scale with all of the consumers right now. So consumers are putting solar panels on their rooftops and large C&I customers are investing in their own generation assets or are going out to the markets to sign up to PPAs.

It's highly unlikely that Microsoft or your neighbour with solar panels will be consuming all of their own energy and nothing but their own energy. So the challenge for energy retailers is exponentially increasing because they're left with the residual consumption or production minus the energy that they consumed themselves. So we can only hope that consumers will maximally use their own consumption or also invest in batteries, even though they're still very expensive. The challenge for energy retailers is that the excess capacity is what they need to manage - and that's much harder than what they used to do, which was just to sell them a relatively predictable energy load. 

And so it feels like the problem is becoming bigger and you will need to use data, you will need to use AI models again, and then it becomes a vicious circle. But you will need to use machine learning models to predict as accurately as possible the individual generation and consumption profiles of everything from B2C to I&C customers with much more data than has ever been used before.

Joris: And I think it's also going to be more and more important to have all of those different forecasts aligned across the retail business as well. There are operational impacts where they have to report to distributors, but there's also a financial impact: what are you then going to buy? What prices are going where? So yes, there will be a big leap in what needs to be forecast because there are just more assets, and more individual or lower level costs that are needed to know what is being produced, and what is being consumed. It's also a question of how retailers can make sure that it's aligned across the entire business so that everything that's happening on the billing side is correct and they can inform the consumers correctly, but also at the higher level they can buy from the wholesale market correctly. So it's going to be a large investment to make all of this match at the right levels and come out with the right prices but also the right reporting back into the market as well.