Edge Labs AI

Intelligence for the physical world.

Edge Labs AI builds systems that let industrial sites understand themselves in real time, on the site itself, without depending on the cloud.

What we build Get in touch

The problem

The world’s largest industries are running blind.

Construction, energy and mining are worth more than seventeen trillion dollars a year. They build the roads, the power, the housing and the raw materials that everything else depends on. And most of them still check their own work by hand, one inspection at a time.

A manager on a major project is often making decisions from a walk-around done last week, covering a small fraction of the site. In the gap between those checks, the difference between what was designed and what is actually being built grows quietly. By the time anyone notices, it is buried under concrete, welded shut, or someone has been hurt.

That gap is expensive. Rework is estimated to waste hundreds of billions of pounds worldwide every year, and these remain three of the most dangerous industries in the world to work in.

~£800bn

Estimated global cost of rework each year

8 in 10

Major projects that finish over budget

1 in 5

Share of US workplace deaths that occur in construction

Sources: citations to be added.

What we do

We give a site a way to see itself.

Edge Labs AI builds two things that work as one system.

The software

SYNAPSE

It turns what a site’s cameras and sensors pick up into a clear, current picture of what is actually there, and holds that against what was supposed to be there.

The hardware

HELIOS

Small, sealed, rugged computers that run SYNAPSE on the site itself, built for heat, dust, vibration and places with no reliable internet.

Together they do something simple to describe and hard to do. They notice, continuously, when reality and the plan stop matching, and they tell the right person while it is still cheap to fix.

Always on.

Not a weekly inspection. A continuous view of the whole site, day and night.

On the site, not in the cloud.

Everything is processed where it happens. Your data does not have to leave your site, and the system keeps working when the connection does not.

People stay in charge.

Every alert can be explained. Every action stays inside limits your team sets. Everything is recorded.

What we build

Two products. One system.

Most companies in this space sell one half of the answer. Either software that assumes a good internet connection, or hardware with nothing intelligent running on it. Industrial sites need both, working together, in conditions neither was designed for. So we build both.

SYNAPSEThe software

SYNAPSE is the software that understands the site.

SYNAPSE takes in what a site’s sensors see, and turns it into a live model of the site as it actually is. It holds that model against the design, the programme and the safety rules, and it looks for the difference.

Perceive. It reads the raw signals coming off the site.
Understand. It works out what those signals mean in context: which element, which stage, which risk.
Act. When something is wrong, it raises it, routes it to the person responsible, and where permitted, triggers a response inside safety limits your team has set.
Evolve. Every deployment makes the system better at the work, without the underlying site data ever having to leave the site.

HELIOSThe hardware

HELIOS is the computer that survives the site.

HELIOS is a family of sealed, ruggedised computers designed to run demanding AI workloads in the field: on battery when the power goes, through dust, heat and vibration, and in hazardous areas where ordinary electronics are not allowed.

They come in several sizes, from a unit one engineer can carry to a cabinet that covers an entire site. They all run the same software. They all work together.

Industries

We start where the consequences are highest.

Construction

The gap between the drawing and the building.

On a large project, thousands of elements are installed every week and only a fraction are ever checked. A column a few centimetres out of position costs almost nothing to correct on the day, and a fortune to correct three floors later. We watch the work as it happens and flag the difference between design and build while it is still a small problem.

Oil and gas

High consequence, low connectivity.

Refineries, terminals and offshore platforms are among the most tightly regulated environments in the world, and among the hardest to monitor. Bandwidth is limited. Data often cannot leave the site for legal reasons. Cloud-first tools were never designed for any of it. Our systems are built to run entirely on site, inside those constraints.

Mining

Too big to walk.

An open pit can be several kilometres across, and underground operations lose satellite positioning altogether. No inspection team can cover ground like that. Continuous, automated observation is the only realistic way to know the state of a site that size, and to see a hazard forming before it becomes an incident.

Beyond these three

The same capability applies wherever a physical environment is high consequence, tightly regulated and poorly connected. We are approaching those markets deliberately, through partners, and in sequence rather than all at once.

Approach

Three commitments we design against.

The intelligence goes where the work is.

Cloud AI assumes a good connection and permission to move your data. Industrial sites usually have neither. We put the computing on the site, so the system works when the network does not, and your data stays yours.

Hardware and software, designed together.

You cannot bolt clever software onto equipment that was never built for the environment, and rugged hardware with nothing intelligent on it is just a box. We design both, so each is built around the other.

The system advises. People decide.

Nothing we build takes a consequential action outside limits a human has set. Every alert can be traced back to what the system saw and why it reached the conclusion it did. Every action is recorded.

Company

Building the intelligence layer for the physical economy.

Edge Labs AI is a deep technology company building the intelligence layer for the physical economy. We are incorporated in the United Kingdom and founded by a team drawn from architecture and construction, artificial intelligence, and industrial operations. We build for the industries where getting it wrong costs money, time and lives.

Mission

To make industrial sites safer, more precise and more efficient by giving the people who run them a real-time picture of what is actually happening, under human oversight, with evidence behind every claim.

Vision

A world where every project, large or small, runs on foresight rather than hindsight.

Safety first.

Every decision is measured against its effect on the people on site.

Precision.

Approximately right is not right.

Honesty.

We do not oversell. Not to customers, not to investors, not to ourselves.

Partnership.

We deploy alongside our customers, not at them.

Careers

We are a small team building difficult things, and we hire people who want that.

If you work in 3D perception, edge systems, embedded engineering or robotics, and would want to join us to work and contribute to this revolutionary endeavour, write to us. Tell us what you have built.

founders@edgelabsai.tech

Building something in a difficult place? We would like to hear about it.

Get in touch

Contact

Tell us about your site.

Whether you run a project or a facility, we are happy to have a direct conversation.

Or email founders@edgelabsai.tech