Why the next defining category of enterprise software will be built on the world itself
By Connor Offutt, Founder & CEO, Aetos
The world today
Here is a fact so obvious that nobody says it out loud: The software that runs the industrial world has barely seen it.
Hang with me.
An ERP system has never seen its plant. The maintenance system has never seen the pump it schedules. Training systems have never seen the machine it certifies people to operate. The system of record has never met a single thing it records.
Our world runs on systems built on a description of reality. Every row in every database is a claim that something exists. Asset is installed. Work order is closed. Employee is certified. And on top of that is a web of siloed point solutions that IT tries to hold together in a giant game of cloud twister.
Each claim begins to drift the moment it’s punched into the system.
Walk any industrial facility with its own asset registry in hand and count the mismatches. The realities don’t match.
Or even better, ask a reliability engineer how much of the week is spent going to look for information the system should already know. Ask why the answer to, “Where’s the isolation procedure?” is still answered with, “In the O&M binder.”
I call the distance between what the records say and what is actually true on the floor the Reality Gap.
And it may be the largest unpriced tax in the industry.
Fig. 1 — What the records say, and what is actually true on the floor.
The Reality Gap is paid in walkdowns and second trips. In tribal knowledge that retires on a Friday. In deferred maintenance nobody can see until it breaks. In near-misses that trace back, not to a careless worker, but to a description that lied.
Historically, this Reality Gap has persisted for one structural reason. There was only one way to update the record: reality, to eyeball, to clipboard, to keyboard, to database. This archaic process is slow, expensive, lossy, and even dangerous. So enterprises did the rational thing and kept record descriptions sparse. Simplified text fields stood in for truth.
The software running the world has been operating on hearsay.
What comes next
We have entered the age of super-intelligence, and the constraints on what is possible are breaking in real time.
The cost of capturing the physical world – turning a place into rich, spatial, machine-readable data – is collapsing by orders of magnitude. What started as a survey crew has become a single operator with a camera, and capturing reality is about to become as ordinary as taking a photograph.
Fig. 2 — The cost of capturing the physical world, collapsing by orders of magnitude.
History is unambiguous about what happens when an input cost collapses a hundredfold. You don’t get better versions of the old workflows. You get inverted architectures.
Cheap compute didn’t give us faster mainframes. It gave us the cloud. Cheap bandwidth didn’t give us better fax machines. It gave us the internet economy. Cheap reality data will not give us a better system of record. It opens the aperture to something completely new.
The inversion is the whole game. Reality stops being what the system describes, and becomes what the system is.
The record becomes an annotation on reality, not the other way around.
The enterprise stack has a vernacular every executive already knows. “The system of record”; “The system of action”; “The system of intelligence.”... Each wave crowned defining companies, and reorganized how the enterprise connects its data with its people.
I believe there is a new wave coming – a new evolution – built on the substrate of truth.
The System of Reality.
We’ve arrived at the junction where software can finally meet the real world.
A System of Reality enables an asset registry that audits itself against the plant floor. A maintenance system that has actually seen the equipment it schedules. A training system that shows, instead of tells.
These were never supposed to be three products. They were one thing – reality – shattered into silos only because software couldn’t see.
The iPhone made phone calls a feature. The System of Reality makes the CMMS a view.
For the first time in history, AI can now effectively reason over 3D space. And this shouldn’t come as a surprise, given that the same scaling laws that have yielded AI chat, image generation, and video are now being applied to the physical world.
The companies that hold the context layer of the physical world will close the Reality Gap. Everything that senses, decides, or acts in industry will be audited and enriched against the primal source of truth: reality itself.
Here is the part I care about most.
America is attempting its largest industrial buildout in generations. New energy, new fabs, new datacenters, and reshored manufacturing are all surging – at the exact moment our most experienced operators are retiring. The expertise is walking out the door faster than it can be replaced. And we cannot rebuild an industrial base faster than we can match the skilled labor needed to keep everything running.
Unless knowledge stops living in binders, and starts living in the place itself.
Fig. 3 — Illustrative. The buildout is surging at the exact moment the expertise is walking out the door.
When the friction between knowing and doing in the physical world approaches zero, industrial productivity compounds. That is the quiet math behind every promise of abundance, starting on the plant floor, not in a slide deck.
This is why we are building Aetos. To close the reality gap, and to give every operator a visual interface that connects them to their workspace. Aetos is a System of Reality.
But this category is bigger than any one company, including mine. There is room in it for the builders, the operators, the visionaries, and the organizations who see the same thing coming.
A category of one is just a product. A category of many is a movement.
The era of software that describes the world is ending.
The era of software that is built on the world has begun.