Google is doubling down on Earth AI, its geospatial platform that could help utilities and critical infrastructure operators plan their resiliency investments.
Late last week, the company announced it has augmented the platform with an AI agent. Initially unveiled in July, Earth AI consists of a vast collection of geospatial models and datasets, including flood forecasting, wildfire detection, and extreme weather models. Users are now able to navigate through them and answer complex questions using a geospatial reasoning agent, powered by Google’s AI engine Gemini.
The platform could be an important addition to the growing array of AI models that utilities are using to protect their infrastructure from extreme weather events, which include wildfire risk and flood monitoring products developed by companies such as Technosylva, Pano AI, and Neara. And in addition to being used by the utilities themselves, it has the potential to augment these companies’ proprietary models and data, as Google’s announcement suggests.
Jim Kapsis, founder of advisory firm The Ad Hoc Group, notes that while Earth AI’s impact and value will depend on how it’s leveraged by utilities and remains to be seen, it is a promising development.
“Utilities are currently inundated with immense amounts of data, and one of their biggest challenges is finding ways to analyze that data and turn it into actionable intelligence,” he told Latitude Media in an email. “Applications like Earth AI, if applied to the utility industry’s unique context, could be transformational.”
A utility looking to harden its infrastructure in the face of an oncoming storm, for example, can use a natural language query to ask the AI agent to help assess its vulnerability. Acting “as an intelligent orchestrator,” the agent provides an answer pulling information from different Earth AI models. In this case, it would use an environmental model to identify the potential path of strong winds and satellite imagery to locate critical infrastructure, as well as population density information pulled from Data Commons, Earth AI’s underlying library of public data.
In a video accompanying Google’s announcement, Nicolas Stussi, managing director of the North America Geospatial Division at Airbus Defence and Space, one of Google Earth AI partners, noted that a similar analysis would have once taken weeks of processing. Now, “you can really do that in a matter of hours,” he said.

The insurance industry and natural disaster recovery are also among the sectors set to benefit from Earth AI’s capabilities, according to Google’s announcement.
Google X moonshot Bellwether, for example, is currently using the platform for its insurance clients, helping them speed up claim payments in the aftermath of an extreme weather event. Similarly, the United Nations Global Pulse, the UN’s innovation and data science initiative, uses them to “assess damage after natural disasters, enabling governments and international organizations to rapidly respond to crises.”
Beyond individual use cases, Kapsis sees Google’s investment in such a tool as an acknowledgment of the consequences of climate change.
“More than anything, it shows that Google is making a bet that the growing risks from extreme weather are so big and touch so many people that it must be involved in providing solutions,” Kapsis said, though he added that this doesn’t come without risks.
“If users, companies, and governments begin to rely on Google’s AI models for emergency preparedness and response, that becomes very powerful and sticky,” he said. “Wouldn’t you trust a company more if their products made you feel safer in what is an increasingly dynamic and dangerous environment?”


