Physical AI research · Edinburgh

Physical intelligencefrom first principles.

∇E → 0 · settling into stable memories

We are building machines that understand the physical world the way humans do, by remembering what they have seen and reasoning about what comes next.

Status

Active development

Location

Edinburgh

Launch

2026

Access

Open inquiry

01Three pillars

See. Reason. Predict.

The loop that connects perception to action.

See · Reason · Predict

01

See

3D understanding from ordinary video.

02

Reason

Physical properties. Uncertainty. Context.

03

Predict

What the world does next, before it does it.

02Mission

At the intersection of 3D perception, physical reasoning, and associative memory.

Built for robotics. Useful everywhere.

Perception that sees in 3D, reasoning that respects physics, and memory that recalls what mattered: the ingredients machines need to operate confidently in the physical world.

3D perception

Scene understanding from ordinary video: geometry, objects, and spatial layout.

Physical reasoning

Properties, forces, uncertainty, and context that govern how the world behaves.

Associative memory

Stable recall and retrieval, patterns that settle like Hopfield networks into memory.

03Namesake

Why we are called Hopfield

Energy landscapes, stable memories, and the physics of learning.

In 1982, John J. Hopfield described a network whose dynamics flow downhill across an energy landscape, settling into stable states that behave like memories. Noisy input is drawn toward the nearest stored pattern. Associative memory as the relaxation of an energy function.

That picture of attractors, basins, and slow descent toward a minimum still echoes in modern attention and retrieval. We named the lab in his honour, and the energy field across this site is a small homage: systems settling, like memory, toward their stable state.

John J. Hopfield · Nobel Prize in Physics, 2024

04FAQ

What to expect

01

What does Hopfield Labs do?

We are building foundational AI infrastructure for machines that need to reason about the physical world. More soon.

02

Who is this for?

Researchers, engineers, and teams building systems that operate in the real world: robotics, simulation, perception, planning.

03

When do you launch?

We are in active development. Get in touch and you will be among the first to know.

05Connect

Stay in touch

Early access, partnerships, or a quick hello.

Topic