👐 As part of our efforts to build an open, synthetic energy data community with @LFE_Foundation OpenSynth - we have also released an updated dataset that contains 10 million synthetic smart meter profiles which you can find on the OS data repository:

🚪 How can you get access to Faraday? Our model is open access. You can sign up directly by visiting

4️⃣ In the previous version of Faraday, we ensured training data was nationally representative for England and Wales. This now also includes Scotland.

3️⃣ As a result of these updates, our model captures low carbon technology profiles more accurately e.g. heat pump loads.

2️⃣ We’ve trained the model on more and updated data, spanning March 2024 to March 2025. This version is trained on data from over one million households.

What can you expect from the new version of our model? 1️⃣ Geographic factors impact household energy consumption. The latest version of Faraday allows users to input location, with outputs reflecting regional - rather than national - consumption patterns.

🚨 We’ve updated our open access generative AI model, Faraday, which outputs synthetic smart meter data ❓ This latest version has locational awareness AND is trained on even more of the latest @OctopusEnergy customer data 🚪 Want access? You can sign up directly - check out

Check out our key recommendations 👇

🌍 These findings show how renewable-powered data centres could mitigate the cost, time and risks that hyperscalers are seeking to avoid. Renewables are a proven and deployable option that can contribute to quickly meeting the demands of AI - and the broader energy transition.

4️⃣ Our model outputs a microgrid that meets consistent yearly data centre demand without needing the grid. To further minimise risk to data centre operators, no/low-utilisation grid ties can provide additional microgrid reliability.

3️⃣Renewable microgrids could be deployed in around half the time of nuclear SMRs (~5 years vs 10 years) and with much greater certainty.

2️⃣It is cheaper by almost a third (31.7%) to power a data centre with 95% renewables (in line with the CP30 target) vs a nuclear SMR.

Headline findings: 1️⃣It is cheaper by approx 43% to power a data centre with a largely renewable microgrid vs a nuclear SMR. Renewables can meet the majority of the constant demands of a large data centre - but gas is still required as a stopgap source of power today, with

💡 Can the UK leverage renewables to power data centres at speed and low cost? 🆕 Our latest analysis compares the cost & speed to power data centres in the UK by leveraging renewable microgrids versus nuclear small modular reactors Check it out 🔗 + 🧵 https://t.co/BEdZ9eOxaM

Day 3 #LSEEnvironmentWeek final three-paper session by @AllanHsiao, Hayeon Jeong & Louise Bernard @CentreNetZero 1️⃣ Political conservation in Brazil & Indonesia 🌴 2️⃣ EV charging response to price 📉in the UK ⚡️ 3️⃣ Congestion zone policy compliance/avoidance in South Korea 🚗

🙏 Thanks to the King Climate Action Initiative @JPAL for funding the trial, @OctopusEnergy for the partnership opportunity, and to everyone who contributed @RDMetcalfe @a_r_schein

🌍 AI can be good for consumers and the grid - helping households save money, integrating more renewable supply, and making the power system more resilient.

💡AI-managed charging can... 1️⃣ Provide significant load-shifting, reducing peak household electricity by 42%. 2️⃣ Work for consumers, with manual overrides accounting for 2.3% of total electricity consumption. 3️⃣ Cut bills by £343/year - or £650 when compared to a standard

🚗 Can AI make EV charging cheaper and greener? CNZ ran the world’s largest AI-managed EV charging trial with 13,000 UK households. Check out the results 👇 Summary & working paper 🔗 https://t.co/G4bYQ5MMtA

🔗 Have a read: