Looking for a startup idea to combat climate change? Consider becoming an expert in home energy assessments. That’s exactly what the founders of French startup Kelvin did, utilizing computer vision and machine learning technology to simplify the process of auditing homes for energy efficiency.
The trio of Clémentine Lalande, Pierre Joly, and Guillaume Sempé turned their attention to home energy audits because they recognized the significant impact that renovations could have on reducing energy consumption and CO 2 emissions. However, like many other sectors in the construction industry, this particular area has yet to fully utilize technology to enhance its processes.
“In Europe, there are 300 million homes in need of renovations within the next 30 years,” explains Lalande, the CEO of Kelvin. “But the construction industry is one of the least digitized sectors, following closely behind agriculture.”
In France, the National Housing Agency (ANAH) has set a lofty goal of 200,000 renovated homes by 2024. But due to the traditional methods used, contractors are falling behind, resulting in a negative impact on the environment.
The regulatory landscape in Europe is favorable for startups focused on this sector. And with its launch in October 2023, Kelvin is purely a software company. Unlike Enter, a similar startup in Germany that we’ve covered, Kelvin does not want to create a marketplace for service providers. Also, unlike Enter, this French startup is not aiming for a customer-facing product.
Instead, the team at Kelvin has developed its own AI model that specializes in home energy assessments through machine learning. The company utilizes a combination of open data, such as satellite images, and its own training data set, which consists of millions of photos and energy assessments.
“We analyze over 12 proprietary, semi-public, and open data sources that provide insights into a building’s thermal performance. Our approach involves utilizing standard segmentation techniques, using machine learning models to identify specific features, such as neighboring buildings, solar panels, and shared ventilation units,” Lalande explains.
“In addition to this, we also have a remote inspection tool that utilizes a bot to direct whoever is on-site to capture specific photos and videos. This valuable data is then used by our models to count radiators, detect doors and ceilings, and even determine the type of boiler or ventilation unit present,” she adds.
Kelvin recognized that in order to establish itself as a widely-used tool for audits, it needed to be accessible on various devices. This is why the startup chose not to utilize LiDAR technology, a 3D scanning method that requires a specific smartphone with a LiDAR sensor.
The startup’s potential clients include construction companies, those in the real estate industry, and even financial institutions seeking to finance home renovation projects. For these organizations, having accurate assessments is essential before making decisions.
In initial tests, the results of Kelvin’s home energy assessments have been within a 5% margin of error of traditional assessments. With widespread adoption of this tool, it will become much easier to compare homes and renovations to make informed decisions.
The startup recently raised €4.7 million (equivalent to $5.1 million) in a round led by Racine², with additional funding from Bpifrance. Seedcamp, Raise Capital, Kima Ventures, Motier Ventures, and several business angels also participated in the investment.
