AlphaEarth Foundations (AEF)

Closed weights Google DeepMind,Google 480M parameters July 2025

No estimate

No hardware requirements for this model

The weights for this model have not been published, so it cannot be downloaded or run on your own hardware at any size. It is reachable only through its provider, and no graphics card changes that.

On record

Full specification

Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.

Origin

Who built this model, where, and when it was published.

Organisation
Google DeepMind,Google
Organisation type
Industry,Industry
Country
United States of America
Published
30 July 2025
Authors
Christopher F. Brown, Michal R. Kazmierski, Valerie J. Pasquarella, William J. Rucklidge, Masha Samsikova, Chenhui Zhang, Evan Shelhamer, Estefania Lahera, Olivia Wiles, Simon Ilyushchenko, Noel Gorelick, Lihui Lydia Zhang, Sophia Alj, Emily Schechter, Sean Askay, Oliver Guinan, Rebecca Moore, Alexis Boukouvalas, Pushmeet Kohli

What it does

The problem areas the model was built for. A model can carry several of each.

Domain
Earth science
Task
Entity embedding, Crop Mapping / Segmentation

Size

How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.

Parameters
480M

"We trained ∼1B and ∼480M parameter variants of AEF, and ultimately proceeded with the smaller variant for improved inference efficiency."

Training data
tokens

[videoframes] "AEF was trained over 8,412,511 video sequences containing interleaved, time-stamped frames from the sources and metadata listed in supplemental materials S1. Each frame covered a 1.28 km x 1.28 km (128 x 128 pixel) area projected into the UTM zone of the area’s centroid and were not limited in length: all available data was used totalling 3,047,520,515 frames." "AEF was trained for 56 hours on 512 TPU v4 devices over 100k steps in batches of 256 video sequences. "

Training compute

The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.

Training compute
8.5 × 10²¹ FLOP

275000000000000 FLOP / sec / TPU v4 * 512 TPUs * 56 hours * 3600 sec / hour * 0.3 [assumed utilization] = 8.515584E21 FLOP

How it was established
Hardware

The training run

What it physically took to train: which chips, how many, for how long, and what that drew from the wall.

Training hardware
Google TPU v4
Chips used
512
Wall-clock time
56 hours

"AEF was trained for 56 hours on 512 TPU v4 devices over 100k steps in batches of 256 video sequences"

Power draw
340.5 kW

Availability

Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.

Weights
Closed — provider access only
Model access
Hosted access (no API)
Training code
Unreleased

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Record confidence
Confident

Sources

Where this record came from and when it was last checked.

Reference
AlphaEarth Foundations: An embedding field model for accurate and efficient global mapping from sparse label data
Last updated
29 May 2026

What the numbers mean

Background

AlphaEarth Foundations (AEF) was published by Google DeepMind,Google, in the country recorded as United States of America, during July 2025. It comes out of an organisation categorised as industry,Industry.

It works in the domain of Earth science, and is recorded as performing the task of entity embedding, Crop Mapping / Segmentation.

Its weights were never published, so it can only be reached through its provider. No graphics card changes that.

What went into building it

The training run consumed about 8.5 × 10²¹ FLOP, on hardware recorded as Google TPU v4. That figure measures what producing the model cost, and has no bearing on how fast it answers.

Answers

AlphaEarth Foundations (AEF) — common questions

01

AlphaEarth Foundations (AEF)— when was it released?

It was published in July 2025.

02

AlphaEarth Foundations (AEF)— what is it used for?

It works in the domain of Earth science, and is recorded as handling the task of entity embedding, Crop Mapping / Segmentation. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

03

AlphaEarth Foundations (AEF)— how much compute was used to train it?

Training consumed around 8.5 × 10²¹ FLOP, on hardware recorded as Google TPU v4. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.

04

AlphaEarth Foundations (AEF)— what GPU do I need to run it?

None. This is a closed model — its weights were never published, so it cannot be downloaded or run on your own hardware at any price. It is reachable only through its provider.

05

AlphaEarth Foundations (AEF)— is it open source?

No. Its weights have not been published, so it exists only as a service controlled by its owner.

06

AlphaEarth Foundations (AEF)— how many parameters does it have?

It has a parameter count of 480M. "We trained ∼1B and ∼480M parameter variants of AEF, and ultimately proceeded with the smaller variant for improved inference efficiency.". That figure is the total, and it is what decides how much memory the model needs — roughly half a gigabyte per billion at the compression most people use.

07

AlphaEarth Foundations (AEF)— who created it?

It was published by Google DeepMind,Google, based in United States of America, an organisation categorised as industry,Industry.

Source

Original publication

Record last updated 29 May 2026

The other direction

Looking at it from the other side?

This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.