Skip to content

Modelling the impact of low carbon technology adoption on energy consumption using generative AI

Our relationship with Octopus Energy means we have access to one of the largest UK smart meter datasets. Ten of thousands of Octopus Energy customers own some form of low carbon technology, such as electric vehicles, heat pumps, solar PV and home batteries. How these customers currently behave can provide important insights for those interested in the energy system of the future. We’re sharing the utility of these insights via a generative AI model called Faraday. 

Partners we’ve worked with in the past have highlighted the value of our smart meter dataset in advancing research, helping policymakers and shaping the Distribution Network Operator (DNO) to Distribution System Operator (DSO) transition. We therefore wanted to unlock its utility for the energy community, whilst strictly protecting customer data and privacy.

A generative AI model trained on Octopus Energy's vast smart meter dataset

Faraday generates daily load “profiles” consisting of half-hourly kWh consumption for a given set of user-specified inputs (e.g. low carbon technology, property type, season etc) which specify a consumer archetype and day type.

Users are able to select from a variety of consumer archetypes, thereby simulating the entire distribution of load profiles of that population instead of a point estimate.

Household smart meter data is a type of personal data, and thus protected by GDPR. Our modelling approach means that we’re able to generate realistic profiles for each consumer archetype that can’t be attributed to individuals.


By modelling at a household level, we can capture the variability of how each household uses energy differently. Researchers, policy makers and DNOs are exploring multiple downstream applications of the tool. Many testers are simulating how households consume energy and model ‘what-if’ scenarios:


What if...

We designed a new innovative tariff

How would this impact energy consumption and the grid?

X% of households in the UK now have electric vehicles

How would it affect energy demand and the grid?

We build more properties of a given type in an area

Could the grid handle it?

Extreme weather conditions like heat waves and cold snaps become more common and severe

What needs to happen to our grid to manage these demand spikes?


Progress to date

our testers

We began development in June 2022 and have built a web app and an API. We have a select group of testers from academic institutions and industry. They are helping us to develop the tool further whilst improving their understanding of electrification at a local level. 

What's next?

We recently released v3 of Faraday, and are currently working on scaling it up to accept more inputs and generate a lot more samples via a single API call. A further update will be released in Q3 2023.

Get in touch

Interested in finding out more about Faraday or trying it out?  Please contact us on