Skip to content

en_tech_talks session 02: the readout

Centre for Net Zero   •   31 May 2022

Buildings generate nearly 40% of annual global CO2 emissions. How can we use data, machine learning and intelligent algorithms to decarbonise the built environment around the world?

Centre for Net Zero recently launched a new event series, en_tech_talks, designed to bring together the brightest and most progressive players in energy and tech, to envision, co-design and develop the future energy system. 

Each session shines a light on leading energy and tech experts at the forefront of the decarbonisation movement – and for the latest ‘buildings edition’, we invited Tom Anderson, co-Founder at Grid Edge, and double act Dan Williams and Josh Eadie, co-Founders of, to tell us more about how they’re disrupting the way we use energy in the built environment. You can find both sets of slides here.

Tom kicked things off with his keynote entitled ‘Buildings Behaving Better: using predictive data to deliver energy efficiency and flexibility in buildings’. He opened the discussion with a simple but compelling idea – that not all kilowatt hours (kWhs) are born equal. Tom is particularly interested in when energy is consumed. Building operators often express interest in optimising the amount of energy that’s being consumed – but as important is when that amount is being consumed. Why? At some points of the day, the electricity we use comes from clean, green sources (when the sun is shining or the wind is blowing) and at others, it comes from fossil fuel sources. In the future, we know that we need to use more of these intermittent, green sources of energy. So how can we prepare buildings for what lies ahead and help them go green?

With 60% of building emissions coming from heating, ventilation and cooling (HVAC), Tom encouraged the audience to consider this challenge by asking them ‘how much is 1°C of comfort worth?’. The answer inevitably depends on a series of factors, such as what building you’re in, what it’s being used for, what time of day it is and so on. Is 1°C of comfort the same in all scenarios? What about the busiest day of the year – which is probably Christmas Eve for a shopping centre – compared to one minute before closing time when it’s relatively empty? This 1°C of comfort changes with electricity prices throughout the day and throughout the year. These prices are also becoming more volatile.

With this in mind, Tom asked the audience to consider how we might change heating/cooling behaviours and a building’s energy profile. We know that a net zero energy system will make increased use of smart tariffs, technology and automation to incentivise greener behaviour from both residential and commercial consumers. If we’re able to start to build a picture, using data, that tells us how a particular building is going to be using energy in the future, we can begin to understand how an energy profile can be adapted.  

Grid Edge’s solution involves building a predictive digital twin of buildings. By taking operational data and creating a predictive model of day-ahead energy profiles, Grid Edge is able to demonstrate the impact of shifting demand on comfort, carbon and cost. This enables buildings to hit a daily CO2 saving goal. 

Tom rounded the discussion off by highlighting the power of a data value stack. To create one, you need to first collect data and then share it with others. By personalising the data and finding a solution to a problem (e.g. showing someone how to get from A to B on Google Maps and letting them make that journey via an Uber), you’re able to co-create value. If you apply this concept to the energy system, you can collect data via smart meters, share it with suppliers, brokers and DSR services, personalise it through time-of-use tariffs and co-create value through peer-to-peer, prosumer and V2G services. 

He suggested Airbnb as the perfect example of value co-creation: a platform that neatly matches the data of people looking for places to stay with the data of those offering their homes up. The idea of building an equivalent platform for the energy system, whereby we dynamically match demand with supply, left the audience with much to consider. 

Next up, Dan and Josh from took to the stage to discuss leveraging machine learning and the Internet of Things to eliminate wasted energy

The issue they’re confronting is wasted ‘Small Power’ (SP) – and it’s bigger than you might think. SP refers to anything that’s plugged-in to a wall. The energy use and emissions impact of these devices is growing and significant. To give you a sense of scale, there are 400,000 offices in the UK – and if you take a typical commercial building, around 20% of its total energy use is wasted through plugged-devices. If you want to manage them more effectively, you can buy smart plugs and sockets, but it’s a manual process and hard to achieve impact at scale.’s solution uses both hardware and software to automate energy management. It combines smart sockets with a machine learning software that has four distinct tiers to it. Tier 1 identifies the devices plugged in, tier 2 enhances this information with contextual data, tier 3 spots ‘wasted energy’ and tier 4 focuses on preventative maintenance and anomaly detection. 

What does this mean in practice? Tier 1 lets buildings see how many laptops, monitors and heaters they’re using and the energy use, cost and emissions associated with each of these items. Tier 2 reveals the impact of contextual factors on energy usage – for example, who is in the building at that time and what the temperature is. Tier 3 involves automatic and instant energy, greenhouse gas and cost reduction across all devices, such as turning off all non-essential appliances if nobody is detected in a room. Finally, tier 4 involves reducing maintenance fees and equipment failures and improving security. 

Value for measurable energy’s customers is found at scale. A company with 2,000 occupants in its office can save over 50% of its energy costs and reduce its carbon emissions by 75 tonnes at the same time in 12 months. 

Dan and Josh closed the discussion by encouraging the audience to try out their new software via a machine learning demo, whereby people can plug a device into a smart socket and it will automatically detect what the device is, its live power usage and how green the energy it’s using is at that exact point in time. 

Once again, we were delighted to be joined by such a strong turnout of guests from around the UK, from energy innovators to academics and policy wonks, who were left with a huge amount of new information and knowledge to consider. 

Keen to come along to the next en_tech_talks session? We’ll be announcing the next date soon so follow us on LinkedIn and Twitter to ensure you don’t miss out. We’re always on the lookout for great speakers too, so if you or someone you know is interested in sharing an inspiring energy tech keynote with our audience, please fill out this form.