At the recent AI for the Planet event, Google outlined a decade-long initiative to leverage artificial intelligence for global disaster resilience, advancing from crisis information delivery to proactive forecasting and detection. The company’s strategic objective remains eliminating surprises during natural disasters by deploying AI-powered tools across its global ecosystem. Significant advancements have been achieved in hydrological and atmospheric modeling. Google’s Flood Hub now issues river flood forecasts up to seven days in advance and urban flash flood warnings up to 24 hours ahead, covering two billion residents across more than 150 countries. The platform relies on extensive historical reporting data and open-sourced hydrology frameworks to ensure global scalability, including in data-scarce regions. For extreme weather, the WeatherNext 2 model generates high-resolution, hourly global forecasts for wind, precipitation, and pressure within minutes, successfully predicting the trajectory and intensity of major cyclones days ahead. Wildfire management is being transformed through the upcoming FireSat constellation, a collaboration with Muon Space and the Earth Fire Alliance that aims to detect fires as small as five square meters anywhere on Earth with twenty-minute update intervals. Additionally, researchers are applying AI to satellite imagery to map urban building reflectivity, guiding cities in deploying cool roof materials to mitigate extreme heat. Beyond predictive modeling, Google has integrated these insights directly into consumer products to deliver actionable crisis data. The Android Earthquake Alerts System provides seismic warnings seconds before shaking occurs, while Public Alerts networks in over ninety countries amplify official emergency notifications. Extreme heat warnings, enriched with safety protocols from the Global Heat Health Information Network, now reach users in more than one hundred countries. To address complex, multi-layered disaster scenarios, Google consolidated its geospatial and climate datasets into the Earth AI collection, enabling institutions to analyze intersections of imagery, population density, and environmental vulnerability for targeted response planning. Daily crisis information interactions on Google Search and Maps average ten million, underscoring the immediate utility of these platforms. Real-world deployment highlights the operational impact of these technologies. In Nigeria and Bangladesh, humanitarian organizations utilized Google’s flood forecasts to distribute anticipatory cash assistance, allowing communities to evacuate before waters rose. During Hurricane Melissa, the U.S. National Hurricane Center leveraged WeatherNext to project landfall in Jamaica five days in advance, enabling local authorities to issue timely public advisories. Google.org continues to fund disaster recovery initiatives and collaborate with UN agencies, scientists, and first responders to standardize data sharing and optimize resource allocation. As machine learning capabilities mature, Google emphasizes that sustained partnerships will be critical to transforming predictive accuracy into measurable reductions in disaster-related casualties and economic loss. The company reaffirms its commitment to expanding these systems globally, moving steadily toward a future where natural disasters are anticipated rather than caught unawares.