Shifting to Meaningful Earthquake Preparedness with AI and Deep Space Data : Risk & Insurance
Peter L. Miller, CPCU, MS, MBA, is chief executive officer of The Institutes and a member of its Senior Management Team.
For as long as we’ve had earthquakes, we’ve accepted one hard truth: you can’t predict them. You can build structures to withstand them. You can install seismographs to detect where ground shaking occurs the moment they begin. But you cannot see them coming.
The best we’ve managed is a few seconds of warning. That’s not enough time to protect people and property and that is why they remain one of the most devastating and costly natural catastrophes.
A recent Munich Re report estimates that a repeat of the 1906 earthquake in today’s San Francisco Bay Area would produce direct losses of $200 billion or more, a figure that has roughly doubled since the last major estimate, because the concentration of technology and infrastructure in Silicon Valley has grown so dramatically.
The report also said the U.S. Geological Survey puts the probability of a magnitude 6.7 or higher earthquake striking the San Francisco Bay Area within the next 30 years at 63 to 72%.
And yet less than 15% of California homeowners carry earthquake coverage, Munich Re noted.
For as long as we’ve treated earthquakes as unforeseeable — something to survive rather than anticipate — that exposure gap seemed like an unavoidable fact of life. Efforts have been made to improve resilience, and strengthen building codes in earthquake-prone areas, but knowing when a big one would strike is a guessing game.
However, that may no longer be the case.
I recently hosted Itamar Zabari, co-founder and CEO of AstroTeq.ai, on the Predict & Prevent® podcast. What he described challenges one of the most stubborn assumptions in seismology and insurance: that earthquake risk can only be managed after the fact.
AstroTeq uses cosmic radiation data from deep space, combined with satellite imagery and AI-driven machine learning, to forecast earthquakes up to 25 days before they strike. Existing systems, such as those used in California and Japan, issue warnings only seconds ahead of a quake.
The implications of early warning of where and when a quake will hit are immediate and concrete. A factory with days of warning before a strong earthquake can shut down machinery, secure equipment, and evacuate non-essential personnel. An energy company with five days of notice can safely shut down pipelines and move a nuclear facility to safe mode.
Itamar told me these preparation steps can reduce damages by a factor of 1:7 to as high as 1:160, depending on the industry. With $200 billion in potential losses on the table in California alone, the value of acting before the shaking starts is not theoretical.
Days and weeks of advance warning lets insurers pre-position loss adjusters, capture pre-event imagery, and accelerate claims settlement. It also gives policyholders a real opportunity to protect their assets, lives and operations before exposure becomes a loss.
The earth has been warning us of impending quakes all along, Itamar told me. “It’s just whispering through different data channels. You need to listen to the whispers and connect them.”
Listen to my conversation with Itamar in the Predict & Prevent® podcast “New Technology Listens for Early Earthquake Warnings.” &
