AI for materials science · Spun out of DLR

AI that remembers
what science forgets.

Elarix is the AI co-scientist for materials R&D. It reads thousands of papers and your lab data, connects them into a physics-aware knowledge graph, and reasons over it to recommend the next experiment worth running.

scroll
The problem with materials R&D

Scientific judgment is scattered.

Decades of materials knowledge sit fragmented across papers, datasets, lab notebooks and people's heads. No AI can learn from what isn't connected — and no scientist can reuse what can't be found. So every program starts from zero.

Published papers4,218
Internal lab runs318
Preprints & datasets12,470
Notebooks & headsuncounted
Vendor spec sheets
The cost of starting over
00
years

to bring a single new material to market, costing billions in missed revenue.

0%
of programs fail

most R&D time is spent in trial-and-error and prototyping waste.

0%
never scale

of new-to-the-world chemical innovations never reach commercial scale.

Your tools store experiments. AI should tell you what to run next.

Introducing Elarix Labs

An AI-native knowledge layer for materials.
And a co-scientist that reasons over it.

Elarix structures the world's materials science — literature, lab data, simulations — into one physics-aware graph, then applies AI reasoning to plan experiments. Every recommendation comes with a source for every claim.

The product

Three AI layers, one co-scientist.

01 · AI knowledge base

Every paper, every run — one AI-readable base.

Elarix's AI reads thousands of scattered papers and internal runs and normalizes them into Composition → Process → Property — one queryable base where every link is traceable to its source.

physics-awarelow hallucinationtraceable
app.elarixlabs.com
Elarix AI knowledge search showing ranked results for SiO₂ aerogel thermal conductivity
02 · AI experiment planning

The next 2–3 experiments, ranked by AI.

Instead of screening fifteen options, the co-scientist reasons over the graph and puts the two or three highest-confidence next steps in front of you — with predicted gain and the evidence behind each one.

app.elarixlabs.com
Elarix AI experiment planner recommending ranked next runs with predicted improvement
03 · In-silico testing

Simulate before you synthesize.

AI-guided molecular-dynamics and DFT simulation screens candidate materials computationally before anything reaches the bench — physics-aware, and every result stays traceable.

app.elarixlabs.com
Elarix knowledge graph linking composition, process and property
AI at the bench
15 2–3

Test 2–3 options instead of 15.

Up to 0% faster R&D iteration, guided by AI.

Who we are

Spun out of the German Aerospace Center.

Built on DLR-proven AI and materials science, by a team that has lived the R&D bottleneck.

Prakul Pandit

Prakul Pandit

Co-founder · CEO
Dora Tempelaars-Busa

Dora Tempelaars-Busa

Co-founder · COO
Prof. Dr. Ameya Rege

Prof. Dr. Ameya Rege

Co-founder · Chief Scientist
Sugan Kanagasenthinathan

Sugan Kanagasenthinathan

Founding Engineer

Let's build AI for materials together.

Pilots now open

Starting in the sustainable polymers sector.

R&D data partnerships

Partnering with academia to train global AI materials models.

contact@elarixlabs.com