Evadapt

European-made. Help advance European tech sovereignty.

Research in progress · Europe

Autonomous intelligence.
Every batch smarter.

Explore the approach

Currently developing and testing the
foundations on synthetic processes.

We’re building fully autonomous AI that can respond to faults it has never encountered, learn from every batch, and adapt to the process it operates. Our goal: a plug-and-play system that brings this intelligence to any industry—without rebuilding it for every factory.

A research-led system, not a standalone or fine-tuned LLM. Physical models, machine learning, planning and memory work together—with LLMs as one component.

Learning across bioreactor batches

Concept illustration · not experimental results

Three wireframe bioreactors: oxygen falls, a response changes aeration, and a record informs a later batch. Each vessel has the same liquid level.Oxygen fallsOxygen recoversOxygen steadierLess feedMore airIncident record

Dissolved oxygen over time · shared scale

Dashed line = reference · schematic only

Unexpected change: oxygen falls in a synthetic bioreactor.

Batch 1 / Detect

Oxygen starts to fall.

Dissolved oxygen drops during the batch. The current model no longer explains what the sensors observe.

Adaptive response: oxygen recovers in a synthetic bioreactor.

Batch 1 / Adapt

More air. Less feed.

In this example, increase aeration and reduce nutrient feed to help dissolved oxygen recover.

Learning across batches: oxygen steadier in a synthetic bioreactor.

Batch 2 / Apply learning

Recognise it earlier next time.

Use checked knowledge from the previous batch to inform an earlier response. The goal: steadier oxygen.

02 / The research

Grounded in physics.
Built to keep learning.

We’re exploring how AI can help manufacturing processes understand change and learn from experience. We test ideas in synthetic processes, asking what works and where the limits remain.

Explore the research
01

Understand the process

Build a clearer picture of how a process behaves.

02

Respond to change

Explore how to adapt when conditions become unfamiliar.

03

Learn from experience

Investigate how each batch can inform the next.

03 / Building Evadapt

Follow the work.
Not just the promise.

A public view of our questions, ambition and next steps.

  1. Foundation

    Build the foundations

    Explore ideas in synthetic processes.

  2. Proof of concept

    Test the ambition

    Understand what works and what needs to improve.

  3. Future pilots

    Learn with manufacturing teams

    Find a bounded problem with measurable value.

A broad direction for the work, not a report of completed milestones.

Follow the journey

04 / Build with us

A big ambition.
Room for you.

We’re looking for co-founders, manufacturing partners and people building Europe’s AI infrastructure.

Build with us

EU-first technology / Customer-data privacy