Construction is teaching AI how to build. Who owns what it learns?

There is a disconnect between what active construction looks like from the outside and what is actually happening inside the system.

That disconnect is about to matter.

You can look down from a high-rise and see a crane sitting still. You can watch concrete get poured and then see the workers go home. You can see nobody working Saturday night and reasonably ask: why is this taking so long?

I saved a video from a construction technology founder describing almost exactly that origin story. For two years they had watched a construction site from their apartment. Workers went home. Nights and holidays weren't being used. The conclusion was: "We are wasting hours." "We should be going 24/7." "We need to have a fully automated robotic construction site."

There is a difference between being close to a construction site and being close to the problem.

The crane isn't the system.

The workers going home aren't the problem. What you cannot see from the apartment window are the contracts, inspections, engineering, sequencing, procurement, safety requirements, logistics, curing, material constraints, jurisdictional requirements, coordination between trades, and thousands of dependencies being managed by people responsible for actually delivering the work.

People outside construction can easily mistake what they can see for how construction actually works.

A fork in the road.

We are at a critical fork with construction technology. There is more capital, technical capability, and attention entering this industry than I have seen in years. I welcome it.

I am building an AI company because I deeply respect the people who show up every day and build the physical economy. Some put on PPE and walk onto a jobsite. Some climb into trucks. Others keep hospitals, factories, and facilities operating. The physical and mental demands can be extreme.

I love these people, and I am building for them.

There is another reason this matters. Construction is one of America's great economic ladders. The trades remain one of the fastest paths I know for a young person without an expensive degree to build a real career and eventually earn six figures. The door is unusually wide.

Men and women. Immigrants. People rebuilding their lives. People with criminal records. People who simply don't fit neatly into the credential-driven parts of our economy. Get on the right crew. Learn. Produce. Become valuable.

Construction can give someone a place where what they can actually do still matters. That is part of what is at stake when we talk about "disrupting construction."

History gave us a warning.

Construction has seen versions of this before. In the late 1990s and early 2000s, outside capital moved aggressively into the mechanical and electrical trades with a consolidation thesis. Successful local contractors were acquired and assembled into national platforms. The promise was scale: purchasing power, national accounts, centralized systems, professional management, and access to capital.

Some of those benefits were real. Some of the largest consolidation experiments also failed.

That history doesn't prove outside capital is bad. It proves something more useful:

Scale and access do not eliminate the requirement to understand the businesses being scaled.

We have watched technology-driven disruption restructure other markets too. Airbnb changed the relationship between property, hospitality, and local communities. DoorDash changed the relationship between restaurants, delivery workers, and customers. Instacart changed the relationship between grocery retailers, shoppers, and consumers.

Those companies created enormous value and convenience. They also demonstrate that technology disruption does not stop at efficiency. At sufficient scale, it can change how people work. How they earn. Who controls important relationships. Where economic value accumulates. And who becomes dependent on whom.

Venture capital is not the enemy. I may raise venture capital myself. Capital can accelerate technology, talent, and good ideas. But capital also comes with an expectation for exceptional growth.

Growth is an incentive. That matters as well-funded and well-resourced technology companies enter construction. Capital can buy speed. Networks can open doors. An investor, accelerator, or board relationship can put a young technology company in front of a major contractor faster than years of grassroots selling ever could.

But capital can accelerate distribution much faster than it can accelerate lived experience.

Access to construction is not the same as understanding construction. Traditionally, construction has been brutally effective at exposing the difference.

You understand a problem. You solve it. Someone in the field tries it. It works. You earn trust. Then you scale.

Technology can reverse that sequence. Gain access. Deploy. Scale. Learn the industry while inside it.

That doesn't mean people from outside construction shouldn't build construction technology. We need exceptional engineers. We need AI researchers. We need entrepreneurs. We need capital.

But they need construction too. Not simply construction's market. Not simply construction's data. They need deep partnership with the people who understand the work.

The bigger question.

And AI makes that relationship fundamentally different. Because while a technology company is learning construction, its systems are learning from construction too.

Our contracts teach them. Our schedules teach them. Our production history teaches them. Our failures teach them. Our recoveries teach them. Our PMs, superintendents, foremen, engineers, and tradespeople teach them.

Over time, those individual pieces begin to form something much more valuable than the documents themselves.

Construction intelligence.

For years we have asked: who owns our data?

I think the bigger question is becoming: who owns the intelligence created from it?

Every project connected to these systems makes them smarter. Every field decision. Every delay. Every recovery. Every estimate. Every change. Every piece of judgment developed over a thirty-year career.

We don't know what that accumulated intelligence will ultimately look like. It may live in models, agents, knowledge systems, local infrastructure, shared industry systems, or technology that hasn't been invented yet. We don't need to know its final form to establish some principles now.

A more responsible path.

And I think we should go further. The people and companies creating that intelligence should participate in the value it creates. If decades of field experience help make an AI system more capable, that value should not simply disappear upstream into a technology company's balance sheet.

Technology companies should win. Investors should win. Contractors should win. The people building the technology should be rewarded for what they create. And the people whose accumulated experience teaches those systems how the physical world gets built should remain participants in the ownership, governance, and value of what those systems learn.

Imagine what we could do with that intelligence if we get this right.

• Preserve the knowledge of a retiring superintendent instead of watching thirty years of experience walk out the door.
• Give an apprentice access to lessons accumulated across generations.
• Reduce the unnecessary cognitive burden carried by PMs, supers, and foremen responsible for increasingly complicated projects.
• Build safer systems.
• Help people make better decisions.
• Use technology to make experienced builders more capable, and young builders more capable faster.

That is the version of construction AI I want to help build. Not technology looking down at the jobsite asking why everyone went home. Technology standing beside the people doing the work and understanding why.

Because construction is not an inconvenience standing between technology and the future.

Construction builds the future.

The semiconductor fabs making AI chips have to be built. The data centers running the models have to be built. The power infrastructure feeding them has to be built. So do our hospitals, factories, laboratories, water systems, schools, roads, and homes.

Builders make commerce possible. And we are about to teach machines an extraordinary amount about how all of that gets done.

There is enormous opportunity here for contractors, technologists, and investors to build this next generation together. But we should establish the relationship while we still have the opportunity to shape it.