Geo for Good is Google’s annual conference focused on their geospatial and cloud offerings. The 2025 edition of the summit was hosted at both New York and Singapore. I was glad to take part in the Geo for Good 2025 Summit at Singapore that took place from Sept 8-11, 2025 at the Google Singapore office.The ConferenceAs Earth Engine gets more integrated with Google’s Cloud offerings, the focus of the conference has also shifted to interoperability and integration with other cloud services. Google has also invested a lot of resources into their AI offerings – particularly with the new Satellite Embedding dataset and it was at the center stage throughout the conference.Geo for Good Singapore Summit Participants (Photo by Google)I saw the following main themes in this year’s conference:Theme 1: InteroperabilityTheme 2: AI and Satellite Embedding DatasetAll the presentations shared during the summit are available on the Summit website and are linked from the ‘Resources’ section under the description of session. While many of the sessions were recorded, they are not released to the public.Theme 1: InteroperabilityEarth Engine used to be a standalone product that has been slowly integrated into Google’s other cloud offering. The development focus has thus shifted from adding new features into EE to enabling integrations with other cloud services. The conference started with a Keynote address on What’s New in Earth Engine which announced some of the new integrations.BigQuery is Google’s data platform for storing and analyzing structured data. It excels at analyzing very large tabular and vector datasets – something that Earth Engine struggles with. There are now close integrations that make both the systems interoperate nicely.ST_RegionStats() function in BigQuery can now do zonal stats with raster data from Earth Engine.ee.FeatureCollection.loadBigQueryTable() function can read tables directly from BigQueryee.FeatureCollection.runBigQuery() can run SQL queries on data from BigQuery and load the results as a FeatureCollection in EE.Google Cloud Storage (GCS) is Google’s storage platform for hosting files. There are now new integrations that allow Earth Engine to work seamlessly with cloud-native geospatial data formats hosted on GCS.The new functions ee.Image.loadZarrV2Array() and ee.ImageCollection.loadZarrV2Array() allow reading ZARR format data in EE. This is a big deal as many large climate and weather datasets can not be used without expensive conversion and ingestion.What’s New in Earth Engine keynoteCloud Native GeoTIFFs (COGs) are already well supported and you can use COGs stored in GCS in EE without uploading them to EE. The COG-backed EE Assets are now quite mature and the preferred way to get your own data in EE. There was a deep-dive and hands-on session exploring these integrations in the Beyond quota limits: transforming Earth Engine workflows with cloud-native formats technical session. This was by far my favorite session at the summit.Justin Braaten’s Beyond quota limits: transforming Earth Engine workflows with cloud-native formats Many users of Earth Engine also use Desktop GIS software. Instead of building cartography and spatial analysis features in Earth Engine, the team has focused on integrating EE with popular GIS software to unlock these new features. Jeremy Malczyk hosted a technical session Data visualization and interaction where they announced the new connectors with desktop GIS. Jeremy Malczyk at Data visualization and interaction sessionGoogle Earth Engine Plugin for QGIS got a major update which adds a set of Processing algorithms with a no-code user interface to visualize and download data from Earth Engine. I was part of the community team that worked on this update and contributed new tutorials to help people use these new features. If you are a QGIS user, do check out my new tutorials:Downloading Images from Earth Engine: A tutorial showing how to create a Sentinel-2 median composite for a region and download it as a GeoTIFF file.Building a Workflow: A tutorial on using the QGIS Model Designer to build a workflow to download a data layer of Cocoa Probability from the Earth Engine Data Catalog and calculate the percentage of cocoa plantation in a plot boundary.Visualizing Data from Earth Engine: Use the Earth Engine Python API to load a data layer from CMIP6 Climate Projections dataset and visualize it on a globe.There is also a new ArcGIS Pro toolbox for to Google Earth Engine that offers similar functionality to ESRI users.Theme 2: AI and Satellite Embedding DatasetYou cannot go to a tech conference nowadays without being bombarded with news about ‘AI’. This summit was no different. Google, with its DeepMind lab, is all in on AI. There were many new announcements on models, agents and AI-enabled datasets. At the center stage was the new AlphaEarth Foundations Satellite Embedding dataset. This dataset solves many of the recurring challenges in working with earth observation data. I was part of the trusted tester program for this dataset and closely worked with the DeepMind team to give feedback and develop learning materials for the launch.I was invited to present my work in the Effective use of Google’s Satellite Embedding dataset session, where I talked about how this dataset makes many tough problems accessible to remote sensing practitioners.Plenary session on Effective use of Google’s Satellite Embedding dataset My talk on the Satellite Embedding DatasetI also launched a new 2-hour hands-n workshop Satellite Embedding Deep Dive at the summit. This is a gentle introduction to the dataset with step-by-step guide and code examples for GEE users and available on our YouTube channel. This is perfect for folks who already use GEE and want to dive into the applications of embeddings in their work.New free workshop Satellite Embedding Deep DiveTechnical SessionsOver the 3-days of the summit, I attended many technical sessions. I want to highlight a few notable sessions.Google DeepMind has taken up the challenging problem of mapping agricultural infrastructure and crops in India. The team presented a preview of their work in the Anthrokrishi session, which was truly groundbreaking. They are still in the early stages and slowly rolling out access to the Agricultural Understanding platform to trusted testers.Google DeepMind’s new Agricultural landscape understanding (ALU) datasetMy interests are around big data processing and automating spatial analysis – so the Automating Earth Engine with Google Cloud was intriguing. We got to deploy Apache Airflow on GCP and set up a workflow using the EE Python API.HackathonThe last day of the summit was a Jam Session where groups get to work together to build a prototype of an idea. I had pitched an idea to take my Similarity Search with Satellite Embedding Dataset and apply it to detect different features and objects. We had many people participate and got very encouraging results. You can see our Jam Session Presentation to learn more.Our hackathon group during the Jam SessionPresenting our resultsMy biggest takeaway from this year’s summit was that Earth Engine is no longer a standalone service. It exists in the Google Cloud ecosystem, and users should leverage these integrations to harness the full power of the platform.Hope this post gives you a useful perspective on the direction of the Earth Engine platform and insight into the new integrations. If you have specific questions on the sessions I described above or any of the new announcements, do drop a comment below.