Safehub taps building-mounted motion sensors and AI to detect earthquakes

Safehub, whose platform enables businesses to monitor their buildings for signs of earthquakes, today closed a $5 million seed round. The company says it will use the capital to accelerate deployment to Fortune 500 customers as it expands its engineering team.

A recent FEMA study pegged U.S. losses from earthquakes at $4.4 billion per year. (Each year, there are on average about 15 earthquakes with a magnitude of 7 or greater, strong enough to cause damage in the billions and significant loss of life.) In spite of the risk, more than 60% of U.S. small businesses don’t have a formal emergency-response plan and fail to back up their sensitive data offsite.

Safehub aims to close the gap with a real-time, building-specific earthquake damage data-gathering solution. The company employs a combination of sensors, cloud-based analytics, and third-party data to provide building-specific earthquake damage information.

In 2019, Safehub teamed up with the Global Earthquake Model (GEM) foundation to model structural robustness directly from its motion sensors installed within buildings. The company used the data to refine the algorithmic predictions of damage from earthquakes that inform its vulnerability estimates, risk and insurance calculations, and the other planning information it provides to customers.

Safehub

Safehub

Safehub’s cell-connected sensors measure earthquake ground motion and building response, as well as changes in buildings’ natural frequencies. The company uses this information to estimate damage to buildings and related business interruption losses. If an earthquake happens, Safehub sends damage alerts and financial loss estimates via text messages, email alerts, and a web dashboard within minutes of the event.

Safehub isn’t the only company intent on tackling the earthquake detection and risk assessment problem. There’s Grillo, an in-home alarm that claims to provide warnings up to two minutes before an earthquake hits. SkyAlert not only gives an early warning of earthquakes but can also turn off gas and assembly lines. That’s not to mention One Concern, which leverages AI and machine learning to advise fire departments how to plan for earthquakes and respond to them.

Some experts are skeptical about these systems’ accuracy. Last February and August, SkyAlert’s app issued alerts that overestimated the magnitude of earthquakes and caused tens of thousands of people to unnecessarily evacuate their workplaces. In response, the Mexico City government adopted a measure preventing private companies from sending alerts to businesses and residents.

But studies suggest algorithms can be trained to predict earthquakes with reasonable accuracy. Researchers from Google’s AI division and Harvard University created an AI model capable of pinpointing the location of aftershocks up to one year after a major earthquake that they found to be more accurate than a method used to predict aftershocks today. And scientists at Stanford developed an AI system — dubbed Cnn-Rnn Earthquake Detector, or CRED — that can isolate and identify a range of seismic signals from historical and continuous data.

“Our vision is a safer and more resilient world,” Safehub CEO Andy Thompson said in a statement. “In the event of a natural disaster, organizations need timely and detailed information to get ahead of potential downtime and business interruption losses. With our platform, business continuity and emergency response professionals can understand the extent of the problem remotely and prioritize damage assessments with building-specific data. With our new investors, we gain the right strategic partners to scale the impact of our technology and improve the resilience of more organizations.”

Fusion Fund and Ubiquity Ventures co-led the seed round in San Francisco-based Safehub, which saw participation from Bolt, Promus Ventures, Blackhorn Ventures, Maschmeyer Group Ventures, and Team Builder Ventures.

Sign up for Funding Weekly to start your week with VB’s top funding stories.

View original article here Source