About Us

We build the runtime where machines think and act.

embedded ocean GmbH, the Munich software company behind Xentara.

Real-time control and machine intelligence belong in one runtime, not in two systems talking past each other. Standard hardware is sufficient. And what you build next stays your decision, not your vendor's.

Who we are

We are where industrial control meets machine intelligence.

embedded ocean was founded in Munich in 2020 by Michael Schwarz and Robert Schachner. Walk into most plants today and the picture is familiar: proprietary controllers, monolithic software stacks and data trapped in silos, all of it built for an earlier era. Incremental upgrades keep that equipment running, but they will not get it to machines that have to act on their own, in real time. So we build the foundation for that instead. One product, Xentara, built in house. And an engineering team that puts it on real machines, whether that is scoping a brownfield retrofit or running an ONNX- model inside a control loop for machine learning. We are a small multinational team with deep industrial engineering experience and no legacy portfolio to protect, which lets us build for where machines are headed instead of the installed base we would otherwise defend.

50+
Customers / Partners
2020
Founded in Munich
9
Nationalities
20+
Engineers
embedded ocean co-founders Michael Schwarz and Robert Schachner
Co-founders Robert Schachner & Michael Schwarz
Where we are going

Bridging the Physical AI Gap

Vision systems, ML models, cloud analytics. Today they all run next to your line, not inside it. Between the data world and the control loop sits a 30 year old architecture that was never designed to let the two meet. That gap is the reason agents stop at the factory gate.

The shift

Industrial control is moving from "read, write, repeat" to "sense, decide, act". A loop that reacts to what it sees needs inference and control on the same clock, not in two systems exchanging messages.

What we are building

One runtime that carries connectivity, deterministic control and AI inference together, on standard hardware, with the machine's semantics built in. The agent asks. Xentara decides. Safety stays enforced by the runtime, never by the model.

Where that leads

Machines that adapt inside the cycle instead of between shifts, on the equipment you already own.

What we stand for

How we build

Innovation through simplicity

Xentara takes on the groundwork so your teams can spend their time on what actually differentiates your machine.

"Just 10% of innovation is actual innovation. The other 90% is laying the groundwork."

Determinism is non- negotiable

Intelligence in the loop is only useful if the loop still holds. Cycle times down to 10 microseconds, deviation measured below 1 microseconds on a 250 microsecond cycle, on an Intel Atom, at 30 percent load, driving a motor over EtherCAT on a single core. If we claim a number, it is a measured one. We deliver proof- not promises.

Simulate before you build

Model based and simulation driven design lets hardware and software teams work in parallel instead of waiting on each other. This eliminates expensive prototyping cycles and bridges the physical & digital thread.

No legacy to defend

We carry no installed base, so every architectural decision serves the machines you will build next, not the ones we sold ten years ago.

Ready to bring Physical AI to the factory floor?

Tell us what your machine has to do that it cannot do today. We will tell you honestly how Xentara can be a fit and what it would take.