GNoME by Google DeepMind: status and timeline

Milestones

Next · Autonomous synthesis of GNoME materials continues

GNoME model developed and trained2023Complete.
Results published in Nature, 2.2M crystals predictedNov 2023Complete.
736 materials independently synthesized2024Complete.
Berkeley A-Lab autonomous synthesis validates predictions2024Complete.

Most important updates

Google DeepMind publishes GNoME: discovered 2.2 million new materials, 736 independently synthesizedNov 2023

Upcoming

Dec 2026Autonomous synthesis of GNoME materials continuesin about 7 weeks (next)

Current obstacles

Synthesis complexity and costPredicting stability is fast, but physically making each material still requires lab work or robotic synthesis, limiting throughput.

Physics limits

Inorganic crystals onlyGNoME focuses on stable inorganic crystal structures. Organic materials and disordered systems are harder to predict with the same approach.
Training data is historicalThe model trains on known materials, so it cannot discover genuinely novel classes of materials outside the feature space of past discoveries.

How it works

3 parts

Neural network

Graph neural networks learn materials relationships

GNoME uses graph neural networks trained on known stable crystals to predict which new compositions and structures are stable.

Prediction

Rank candidates by stability likelihood

The model scores each candidate structure, prioritizing those most likely to be stable and useful, focusing human and robotic effort.

Validation

External synthesis confirms predictions

Predicted materials are synthesized in labs and by autonomous systems. High validation rates show the model learned real materials science.

Spec sheet

Spec	GNoME
Predicted crystals	2.2 million^R (reported)
Top-tier candidates	380,000^R (reported)
External lab synthesis	736^R (reported)
Berkeley A-Lab synthesis	41+ new materials^R (reported)
Equivalent discovery time	~800 years^E (estimated)

^R reported by the company · ^E our estimate

Update log

1 update

Nov 2023

Major: Google DeepMind publishes GNoME: discovered 2.2 million new materials, 736 independently synthesizedBlogResearch

About Google DeepMind

The team behind GNoME

Google DeepMind

Gemini, AlphaFold, Genie, Veo, TPU chips

Google DeepMind is Alphabet's AI lab, formed in 2023 from DeepMind and Google Brain. It builds the Gemini models and leads in AI for science. AlphaFold won a share of the 2024 Nobel Prize in Chemistry.

Gemini 3 (Nov 2025) led many benchmarks. Gemini 3.5 Pro, announced in May 2026, is delayed.
AlphaFold predicted structures for over 200 million proteins and won a share of the 2024 Chemistry Nobel.
Google's own Ironwood TPU chip is used in-house and sold to outside labs, including Anthropic.

deepmind.google (opens deepmind.google)Company profile →

Founded: 201016 yrs
Headquarters: United KingdomUnited States
Status: Subsidiaryof Alphabet
Valuation: Alphabet-owned
Works in: AIFrontier modelsComputingBiotech
Coverage: 11 programs · 40 updateslatest 2d agochecked 25 Sep
Customers: Anthropic
Partners: Isomorphic LabsBroadcom
Acquired by: Alphabet
People: Demis HassabisCEO and co-founderShane LeggCo-founder and Chief AGI ScientistJohn JumperAlphaFold lead; 2024 Nobel laureate

Spec

GNoME

Predicted crystals

2.2 million^R (reported)

Top-tier candidates

380,000^R (reported)

External lab synthesis

736^R (reported)

Berkeley A-Lab synthesis

41+ new materials^R (reported)

Equivalent discovery time

~800 years^E (estimated)