Google Earth AI now connects weather, population, and imagery models through Geospatial Reasoning so you can ask a single question and get a combined answer. Gemini chat in Google Earth is rolling out in the U.S. with higher limits for Google AI Pro/Ultra subscribers and coming to Earth Professional tiers “in the coming weeks.” Pricing for Earth: Standard (free), Professional ($75), Professional Advanced ($150) per user per month.
What is Google Earth AI?
Google Earth AI is Google’s family of geospatial AI models and reasoning agents built to turn planetary data into answers. It sits alongside Google Earth (visualization and lightweight analysis), Google Maps Platform (APIs like Imagery Insights), and Google Cloud (to run models at scale or add your own data).
Under the hood are two pieces worth knowing:
- Geospatial Reasoning: an agent that chains together Earth AI models and data (like forecasts, population, and imagery) to answer multi-part questions such as “Which schools are at risk if this storm hits here?”
- AlphaEarth Foundations & Satellite Embeddings: a DeepMind model that compresses a year of multi-sensor Earth observation into a 64-value vector for each 10 m pixel, available annually from 2017 in Earth Engine. In practice, that means faster change detection and classification without training heavy models from scratch.
What’s new this week
1) Geospatial Reasoning can now “connect the dots.”
Instead of running flood, population, and infrastructure checks separately, Gemini can coordinate those models so you get a combined view of impact and vulnerability in one shot. Google calls this a framework powered by Gemini’s reasoning that chooses the right models, runs them, and explains the result.
2) Gemini chat in Google Earth gets more capable and expands access.
Analysts can ask Earth to “find algae blooms in rivers” or spot dried-up river sections that may trigger dust storms. Google says this experimental capability is rolling out in the U.S., with higher limits for Google AI Pro and Ultra subscribers today, and coming in the next few weeks to Earth Professional/Professional Advanced users. WIRED and Google’s docs clarify Professional pricing at $75 and Professional Advanced at $150 per user per month.
3) Earth AI models on Google Cloud for trusted testers.
Organizations can combine their own datasets with Earth AI Imagery, Population, and Environment models and Google’s Maps Platform data (e.g., Imagery Insights) to solve specific problems like vegetation encroachment or storm-path planning.
Availability and pricing
- Google Earth plans:
- Standard: free.
- Professional: $75/user/month, adds layers like land surface temperature.
- Professional Advanced: $150/user/month, adds EV chargers, tree canopy, and more.
- Gemini chat access today: U.S. users with Google AI Pro/Ultra get higher limits in Google Earth right now; Professional/Pro Advanced users in the U.S. get the new capabilities “in the coming weeks.”
- Trusted Testers on Google Cloud: apply for consideration if you’re an enterprise, city, or nonprofit working on social impact.
Real-world examples (mini case studies)
- WHO AFRO (cholera risk): using Population and Environment models with their data to map where cholera outbreaks are likely, supporting vaccination and water/sanitation decisions.
- Planet & Airbus (imagery providers): applying Earth AI models at scale to detect deforestation or power-line vegetation encroachment, so utilities can prevent outages.
- Bellwether + McGill and Partners (insurance): hurricane-related insights that help insurers pay claims faster after landfall.
How to try it: a short workflow
- Pick your access path
- U.S. subscriber to Google AI Pro/Ultra and want higher chat limits in Earth right now? You’re covered.
- Need professional layers and soon-to-arrive Gemini features in Earth? Professional or Professional Advanced.
- Enterprise or nonprofit with your own data pipelines? Apply for Trusted Tester on Google Cloud.
- Ask a clear question
Example: “Find algae blooms in these counties, then rank the nearest intakes for drinking water risk.” Start broad, then narrow with filters. - Validate the result
Cross-check with local sensor data, historical imagery, and known incident reports. - Export or extend
Hand off results to Earth Engine for deeper analysis or Maps Platform / Cloud for app and dashboard work.
Google Earth AI vs Earth Engine vs traditional GIS
| Task | Google Earth AI | Google Earth Engine | Traditional GIS stack |
|---|---|---|---|
| Natural-language Q&A | Yes, Gemini chat (U.S. availability expanding) | No chat, code/API driven | No |
| Model chaining (weather + population + imagery) | Geospatial Reasoning | Build workflows manually | Manual, multi-tool |
| Data granularity | 10 m embeddings (2017+) where applicable | Petabyte catalog, many resolutions | Varies by provider |
| Pricing | Standard free, Pro $75, Pro Adv $150 | Usage-based on GEE + platform fee | Licenses + infra |
| Best for | Fast “what/where” answers | Heavy analysis and exports | Custom pipelines |
Glossary
- Geospatial Reasoning: An AI agent that picks data/models and executes a plan to answer a geo question.
- AlphaEarth Foundations: A model that turns a year of satellite data into a 64-number vector per 10 m pixel.
- Satellite Embedding dataset: Precomputed AlphaEarth vectors in Earth Engine from 2017 onward.
- Trusted Tester: Early-access program for organizations to try features before general release.
Limitations and considerations
- Coverage and timeliness vary by region and data partner. Confirm local ground truth before acting.
- U.S.-first availability means some features may not appear in your country yet.
- Model explanations are improving, but you should still document assumptions and thresholds for audits.
- Costs scale with users and layers; plan governance and budgets accordingly.
Frequently Asked Questions (FAQs)
Is Google Earth AI the same as Earth Engine?
No. Earth AI focuses on reasoning and answers through chat and model chaining. Earth Engine is a petabyte-scale analysis platform and catalog for heavy processing. They complement each other.
Is this available outside the U.S.?
Some features are U.S.-first. Google says access is expanding; Cloud Trusted Tester programs may admit global orgs case-by-case.
How do I get higher chat limits?
In the U.S., Google AI Pro/Ultra subscribers get higher limits in Google Earth now. Pro/Pro Advanced access is coming in weeks.
What does Professional vs Pro Advanced add?
Professional adds extra data layers; Professional Advanced adds even more (e.g., EV chargers, canopy) and higher design limits. Pricing is $75/$150 per user per month.
Do I need to code?
For chat workflows, not necessarily. For deeper analytics, you may still use Earth Engine scripts or Maps/Cloud APIs.
Can I use my own satellite imagery?
Trusted Testers on Google Cloud can mix their data with Earth AI models and Google datasets.
What are “embeddings” here?
A compact numeric representation of a place and year. AlphaEarth’s 64-dim vectors capture multi-sensor features to speed learning.
How reliable are the answers?
They’re improving, but you should validate outputs with local data. Treat them as decision support, not ground truth.
Featured Snippet Boxes
What’s new in Google Earth AI this week?
Google added Geospatial Reasoning to connect weather, population, and imagery models, and it’s expanding access to Gemini chat in Google Earth. U.S. AI Pro/Ultra subscribers get higher limits now, with Professional tiers gaining the capability in the coming weeks.
How much does Google Earth Professional cost?
Professional costs $75 per user per month. Professional Advanced costs $150. Standard is free. Some new Gemini-powered capabilities are rolling out to these paid plans first in the U.S.
What is AlphaEarth Foundations in plain English?
It’s a way to compress a full year of satellite data for each 10 m pixel into 64 numbers. These “embeddings” speed up change detection and classification without training heavy models each time. Annual layers from 2017 are available in Earth Engine.
Can I bring my own data?
Yes. Trusted Testers can combine their own datasets with Earth AI models on Google Cloud and Maps Platform data like Imagery Insights to build targeted solutions. Apply for consideration.
Who’s using this today?
Examples include WHO AFRO for cholera risk, Planet for deforestation mapping, Airbus for power-line vegetation checks, and Bellwether for hurricane insights that help insurers pay claims faster.
Source: Google Blog