Google's AlphaEarth Foundations: AI Model Revolutionizes Earth Observation and Mapping

Google's Newest AI Model Acts like a Satellite to Track Climate Change

AlphaEarth Foundations is a chip off Google DeepMind block -- and it's here to help save the world. Google's newest AI model is going to scour the Earth and, ideally, help it out. That's the plan, anyway. The mission is to find out once and for all, in fine detail, what we are doing to our planet. Crucially, once the model has supposedly done this it will also, apparently, explain where we might be able to best put things in place to help our world. AlphaEarth Foundations, an offshoot of Google's DeepMind AI model, aims to leverage machine learning and all the gobs and gobs of data that Google has absorbed about our planet over the last two decades, in order to understand how specific areas are changing over time. The model uses a system called "embeddings" that takes terabytes of data collected from satellites every day, analyzes it, and compresses it down to save storage space. The result is a model of different filters overlaid over maps that are color coded to indicate material properties, vegetation types, groundwater sources, and human constructions such as buildings and farms. Google says the system will act as a sort of "virtual satellite," letting users call up on demand detailed information about any given spot on the planet. The goal, Google says, is for users of the service to be able to better understand how specific ecosystems on the planet work, including how air quality, sunlight, groundwater, and even human construction projects vary and change across a landscape. Ultimately, the company wants the model to help answer questions from paying governments and corporations that wish to know, for example, which ecosystems may have more access to sunlight or groundwater that can help determine the best spots to grow a certain crop. Alternatively, it may aid in identifying areas to plop down solar panels with maximum payoff, or build structures in more climate resilient locations. Google's new model has already mapped a complex surface in Antarctica -- an area notoriously difficult to capture due to irregular satellite imaging -- in clear detail. It has also supposedly outlined variations in Canadian agricultural land use that are invisible to the naked eye.

Google calls its new AI model a "virtual satellite.

Called AlphaEarth Foundations, the model stitches together data from actual satellite images, radar, climate simulations, and more to map Earth's land and coastal waters. "The Satellite Embedding dataset is revolutionizing our work by helping countries map uncharted ecosystems - this is crucial for pinpointing where to focus their conservation efforts," Nick Murray, director of the James Cook University Global Ecology Lab and Global Science Lead of Global Ecosystems Atlas, said in a Google DeepMind blog post.

Google's 'virtual satellite' AI model can provide a near real-time view of Earth

It can help researchers track changes to our planet over time. Google has introduced a new AI model called AlphaEarth Foundations that it says can function like a "virtual satellite." The model uses a system called "embedding," which works by taking big volumes of pubic information from various sources every day, such as optical satellites, radars and climate simulations, and then combining them all together. It then divides lands and coastal waters into 10x10 meter squares, which it then analyzes and tracks over time. As Wired explains, these squares are color-coded to indicate different characteristics, such as vegetation types and material properties. The company said AlphaEarth Foundations makes its data easy to use by creating what it calls "highly compact summary" for each square of land or coastal water it monitors. These compact summaries apparently need 16 times less storage compared to those produced by comparable AI systems, thereby reducing costs needed for Earth observation. So what exactly can the model's data be used for? Google explained that scientists can use the model to create detailed maps on demand for multiple purposes, such as to monitor crop health or to track deforestation. In its announcement, the company claimed that the model excelled at a wide range of tasks over different time periods when it was tested. "AlphaEarth Foundations represents a significant step forward in understanding the state and dynamics of our changing planet," Google wrote. The company gave over 50 organizations access to the model's Satellite Embedding dataset, a collection of its annual embeddings, to test its use in real world applications over the past year. Now, it has released the dataset in Google Earth Engine so that other scientists can use it for their own research.

A new Google Earth AI model aims to make satellites obsolete

I tried using Android without any Google apps. Here's how far I could go Google's new AlphaEarth Foundations AI isn't a satellite. It's a system designed to make satellites themselves a lot less indispensable. Instead of waiting for orbital flyovers and stitching together uneven, sometimes cloudy datasets, AlphaEarth Foundations builds a persistent, unified model of the Earth's surface and coastal waters. Think of it as a "virtual satellite" that can see more clearly, more consistently, and in some ways more accurately than traditional space-based sensors (Source: Google Deepmind). A fascinating use of AI This is way more than your grandpa's Google Earth The color-coded, information-dense imaging representing Alphaearth Foundations' advanced mapping. At its core, AlphaEarth Foundations integrates petabytes of Earth observation data -- from radar and optical satellites to 3D lidar mapping and climate models -- into what Google calls a global embedding. This embedding condenses massive datasets into uniform, easy-to-use summaries that describe the planet in 10-by-10-meter squares. The end result is a digital Earth twin that's compact enough to analyze at scale, but rich enough to track changes in land use, vegetation health, and even coastal dynamics. Where satellites stumble, AlphaEarth Foundations steps in. The team behind the model showed off examples in Ecuador, where persistent cloud cover has historically made agricultural monitoring nearly impossible. AlphaEarth was able to cut through the noise, mapping farmland in different stages of development without waiting for clear skies. In Antarctica, a nightmare scenario for most orbiting imagers due to irregular coverage and low light, the model rendered surface features in crisp detail. In Canada, it exposed subtle agricultural land-use changes invisible to conventional satellite imaging. Each embedding consists of 64 dimensions. Taken as a whole, they cover the entire globe. The model doesn't just produce prettier maps. Its technical efficiency matters just as much. According to Google, the embeddings require 16 times less storage than comparable AI systems, making them cheap to use at planetary scale. In head-to-head testing, AlphaEarth Foundations delivered a 24% lower error rate than other leading AI mapping models, even when label data was scarce. That kind of efficiency translates directly into faster, more accurate decision-making for scientists and policymakers. And the tech isn't locked behind a corporate wall. Google is releasing a massive set of AlphaEarth embeddings -- over 1.4 trillion footprints per year -- through the Google Earth Engine as the Satellite Embedding dataset. More than 50 organizations, including the United Nations Food and Agriculture Organization, Stanford University, and Brazil's MapBiomas project, have already put the system to work. Examples of how embeddings take into accout data from various sources and timeframes. The use cases are real and immediate. The Global Ecosystems Atlas is using the dataset to classify minimally or unmapped ecosystems such as hyper-arid deserts and coastal shrublands -- data critical for conservation planning. In Brazil, MapBiomas is leveraging the embeddings to monitor agricultural expansion and deforestation in the Amazon, generating faster and more precise insights than traditional methods. Nick Murray, Director of the Global Ecology Lab at James Cook University, called the dataset "revolutionary" for ecosystem mapping. MapBiomas founder Tasso Azevedo said it enables "maps that are more accurate, precise, and fast to produce, [which is] something we would have never been able to do before." Long term, AlphaEarth could be paired with reasoning LLMs like Google's Gemini to unlock even more advanced geospatial intelligence. For now, though, the system represents a seismic shift in Earth observation: a move from relying on the timing and limits of satellites to a constantly updated, AI-driven view of the planet. If AlphaEarth Foundations delivers on its early promise, the future of global monitoring may not orbit above us -- it may already be here.

AlphaEarth Foundations helps map our planet in unprecedented detail

New AI model integrates petabytes of Earth observation data to generate a unified data representation that revolutionizes global mapping and monitoring Every day, satellites capture information-rich images and measurements, providing scientists and experts with a nearly real-time view of our planet. While this data has been incredibly impactful, its complexity, multimodality and refresh rate creates a new challenge: connecting disparate datasets and making use of them all effectively. Today, we're introducing AlphaEarth Foundations, an artificial intelligence (AI) model that functions like a virtual satellite. It accurately and efficiently characterizes the planet's entire terrestrial land and coastal waters by integrating huge amounts of Earth observation data into a unified digital representation, or "embedding," that computer systems can easily process. This allows the model to provide scientists with a more complete and consistent picture of our planet's evolution, helping them make more informed decisions on critical issues like food security, deforestation, urban expansion, and water resources. To accelerate research and unlock use cases, we are now releasing a collection of AlphaEarth Foundations' annual embeddings as the Satellite Embedding dataset in Google Earth Engine. Over the past year, we've been working with more than 50 organizations to test this dataset on their real-world applications. Our partners are already seeing significant benefits, using the data to better classify unmapped ecosystems, understand agricultural and environmental changes, and greatly increase the accuracy and speed of their mapping work. In this blog, we are excited to highlight some of their feedback and showcase the tangible impact of this new technology.

Google DeepMind says its new AI can map the entire planet with unprecedented accuracy

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now Google DeepMind announced today a breakthrough artificial intelligence system that transforms how organizations analyze Earth's surface, potentially revolutionizing environmental monitoring and resource management for governments, conservation groups, and businesses worldwide. The system, called AlphaEarth Foundations, addresses a critical challenge that has plagued Earth observation for decades: making sense of the overwhelming flood of satellite data streaming down from space. Every day, satellites capture terabytes of images and measurements, but connecting these disparate datasets into actionable intelligence has remained frustratingly difficult. "AlphaEarth Foundations functions like a virtual satellite," the research team writes in their paper. "It accurately and efficiently characterizes the planet's entire terrestrial land and coastal waters by integrating huge amounts of Earth observation data into a unified digital representation." The AI system reduces error rates by approximately 23.9% compared to existing approaches while requiring 16 times less storage space than other AI systems. This combination of accuracy and efficiency could dramatically lower the cost of planetary-scale environmental analysis. The core innovation lies in how AlphaEarth Foundations processes information. Rather than treating each satellite image as a separate piece of data, the system creates what researchers call "embedding fields" -- highly compressed digital summaries that capture the essential characteristics of Earth's surface in 10-meter squares. "The system's key innovation is its ability to create a highly compact summary for each square," the research team explains. "These summaries require 16 times less storage space than those produced by other AI systems that we tested and dramatically reduces the cost of planetary-scale analysis." This compression doesn't sacrifice detail. The system maintains what the researchers describe as "sharp, 10×10 meter" precision while tracking changes over time. For context, that resolution allows organizations to monitor individual city blocks, small agricultural fields, or patches of forest -- critical for applications ranging from urban planning to conservation. More than 50 organizations have been testing the system over the past year, with early results suggesting transformative potential across multiple sectors. In Brazil, MapBiomas uses the technology to understand agricultural and environmental changes across the country, including within the Amazon rainforest. "The Satellite Embedding dataset can transform the way our team works," Tasso Azevedo, founder of MapBiomas, said in a statement. "We now have new options to make maps that are more accurate, precise and fast to produce -- something we would have never been able to do before." The Global Ecosystems Atlas initiative employs the system to create what it calls the first comprehensive resource for mapping the world's ecosystems. The project helps countries classify unmapped regions into categories like coastal shrublands and hyper-arid deserts -- crucial information for conservation planning. "The Satellite Embedding dataset is revolutionizing our work by helping countries map uncharted ecosystems -- this is crucial for pinpointing where to focus their conservation efforts," said Nick Murray, Director of the James Cook University Global Ecology Lab and Global Science Lead of Global Ecosystems Atlas. The research paper reveals sophisticated engineering behind these capabilities. AlphaEarth Foundations processes data from multiple sources -- optical satellite images, radar, 3D laser mapping, climate simulations, and more -- weaving them together into a coherent picture of Earth's surface. What sets the system apart technically is its handling of time. "To the best of our knowledge, AEF is the first EO featurization approach to support continuous time," the researchers note. This means the system can create accurate maps for any specific date range, even interpolating between observations or extrapolating into periods with no direct satellite coverage. The model architecture, dubbed "Space Time Precision" or STP, simultaneously maintains highly localized representations while modeling long-distance relationships through time and space. This allows it to overcome common challenges like cloud cover that often obscures satellite imagery in tropical regions. For technical decision-makers in enterprise and government, AlphaEarth Foundations could fundamentally change how organizations approach geospatial intelligence. The system excels particularly in "sparse data regimes" -- situations where ground-truth information is limited. This addresses a fundamental challenge in Earth observation: while satellites provide global coverage, on-the-ground verification remains expensive and logistically challenging. "High-quality maps depend on high-quality labeled data, yet when working at global scales, a balance must be struck between measurement precision and spatial coverage," the research paper notes. AlphaEarth Foundations' ability to extrapolate accurately from limited ground observations could dramatically reduce the cost of creating detailed maps for large areas. The research demonstrates strong performance across diverse applications, from crop type classification to estimating evapotranspiration rates. In one particularly challenging test involving evapotranspiration -- the process by which water transfers from land to atmosphere -- AlphaEarth Foundations achieved an R² value of 0.58, while all other methods tested produced negative values, indicating they performed worse than simply guessing the average. The announcement places Google at the forefront of what the company calls "Google Earth AI" -- a collection of geospatial models designed to tackle planetary challenges. This includes weather predictions, flood forecasting, and wildfire detection systems that already power features used by millions in Google Search and Maps. "We've spent years building powerful AI models to solve real-world problems," write Yossi Matias, VP & GM of Google Research, and Chris Phillips, VP & GM of Geo, in an accompanying blog post published this morning. "These models already power features used by millions, like flood and wildfire alerts in Search and Maps; they also provide actionable insights through Google Earth, Google Maps Platform and Google Cloud Platform." The release includes the Satellite Embedding dataset, described as "one of the largest of its kind with over 1.4 trillion embedding footprints per year," available through Google Earth Engine. This dataset covers annual snapshots from 2017 through 2024, providing historical context for tracking environmental changes. Google emphasizes that the system operates at a resolution designed for environmental monitoring rather than individual tracking. "The dataset cannot capture individual objects, people, or faces, and is a representation of publicly available data sources, such as meteorological satellites," the company clarifies. The 10-meter resolution, while precise enough for most environmental applications, intentionally limits the ability to identify individual structures or activities -- a design choice that balances utility with privacy protection. The availability of AlphaEarth Foundations through Google Earth Engine could democratize access to sophisticated Earth observation capabilities. Previously, creating detailed maps of large areas required significant computational resources and expertise. Now, organizations can leverage pre-computed embeddings to generate custom maps rapidly. "This breakthrough enables scientists to do something that was impossible until now: create detailed, consistent maps of our world, on-demand," the research team writes. "Whether they are monitoring crop health, tracking deforestation, or observing new construction, they no longer have to rely on a single satellite passing overhead." For enterprises involved in supply chain monitoring, agricultural production, urban planning, or environmental compliance, the technology offers new possibilities for data-driven decision-making. The ability to track changes at 10-meter resolution globally, with annual updates, provides a foundation for applications ranging from verifying sustainable sourcing claims to optimizing agricultural yields. The Satellite Embedding dataset is available now through Google Earth Engine, with AlphaEarth Foundations continuing development as part of Google's broader Earth AI initiative. As one researcher noted during the press briefing, the question facing organizations isn't whether they need planetary-scale intelligence anymore -- it's whether they can afford to operate without it.

Google Earth AI: Our state-of-the-art geospatial AI models

Sorry, your browser doesn't support embedded videos, but don't worry, you can download it and watch it with your favorite video player! We've spent years building powerful AI models to solve real-world problems. Today we're introducing Google Earth AI, our collection of geospatial models and datasets to help people, businesses and organizations tackle the planet's most critical needs. AlphaEarth Foundations, also announced today, is part of Google Earth AI. Google Earth AI expands on our recent Geospatial Reasoning effort and includes models that address multiple areas. Notable examples include detailed weather predictions, flood forecasting and wildfire detection. Other models help improve urban planning and public health by providing a rich understanding of imagery, population dynamics and urban mobility. These models already power features used by millions, like flood and wildfire alerts in Search and Maps; they also provide actionable insights through Google Earth, Google Maps Platform and Google Cloud. As we continue this work, we're committed to giving people the information they need to solve some of the biggest challenges of our time.

DeepMind unveils AlphaEarth to map planet in 10-meter detail

Google DeepMind has announced a new artificial intelligence system, AlphaEarth Foundations, designed to analyze the Earth's surface by integrating vast amounts of observation data into a unified digital model. The system processes information from multiple sources, including optical satellite images, radar, 3D laser mapping, and climate simulations. The core technology creates highly compressed digital summaries, or "embedding fields," for 10×10 meter squares of the planet's terrestrial land and coastal waters. According to the research paper, this approach reduces error rates by approximately 23.9% compared to existing methods while requiring 16 times less storage space than other tested AI systems. A key technical feature is its support for "continuous time," allowing it to generate accurate maps for specific date ranges by interpolating data to fill gaps caused by factors like cloud cover. For the past year, over 50 organizations have been testing the system. In Brazil, the organization MapBiomas is using the technology to monitor agricultural and environmental changes, including deforestation in the Amazon. The Global Ecosystems Atlas initiative is also using it to create a comprehensive resource for mapping the world's ecosystems to aid conservation efforts. The system's "Satellite Embedding dataset," which contains over 1.4 trillion data footprints per year, is now available through Google Earth Engine. This dataset includes annual snapshots from 2017 through 2024. Google states that the 10-meter resolution is designed for environmental monitoring and cannot capture or identify individual people, faces, or objects.