Eric Wu Thinks Construction Workers Need an AI Copilot
Silicon Valley has built AI copilots for programmers, lawyers and doctors. Eric Wu wants to build one for the person holding the drill.
The Opendoor co-founder is back with NavigateAI, a startup building AI tools for construction workers and other field professionals.
NavigateAI launched with $25 million in seed funding at a $225 million post-money valuation, led by Elad Gil, with backing from Khosla Ventures, Fifth Wall, Lennar, Tishman Speyer and others.
Its pitch is surprisingly straightforward.
Point a camera at the job.
Ask a question.
Let AI help determine what needs to happen next.
The product is designed for workers, not managers
Construction technology has traditionally focused heavily on project managers.
Software tracks budgets, timelines, drawings and progress.
NavigateAI is going after a different user: the worker actually installing something.
Its software can run on smartphones and, in hands-free scenarios, through Meta's AI-enabled glasses.
A worker might point the camera at an installation and ask whether something has been fitted correctly, what torque specification is required or whether the work complies with building requirements.
NavigateAI can then pull information from specifications, manufacturer documentation and company policies.
The ambition is essentially to give every worker an experienced expert who is available instantly.
AI’s data-center boom created the problem NavigateAI wants to solve
There is an interesting circularity to the business.
AI is driving enormous investment into data centers.
Those data centers need people to build them.
And the construction industry already has a labor shortage.
The Associated Builders and Contractors estimated the U.S. construction industry would need hundreds of thousands of additional workers in 2026 just to meet demand. Large AI infrastructure projects can require thousands of workers at peak construction.
NavigateAI is betting software can make the available workforce more productive.
Instead of replacing construction workers, this is AI being positioned as a tool for making human workers more capable.
The business model is unusual
NavigateAI initially experimented with usage-based pricing.
But newer contracts are moving toward a model where the company gets paid based on the economic value it helps create.
If its software helps a builder significantly reduce the cost of constructing a home, for example, NavigateAI can receive a percentage of the savings.
That creates obvious upside.
It also creates a measurement problem.
Was a project completed faster because of AI?
Or because the weather was better?
Or because a particular crew was more experienced?
Value-based pricing sounds attractive until everyone has to agree on exactly where the value came from.
The real asset might eventually be the data
There is another layer to NavigateAI's business.
Every worker using the system can potentially generate video showing how physical tasks are performed.
Correct installations.
Incorrect installations.
Tools.
Materials.
Movement.
Repair work.
That could eventually become valuable training data for robotics.
Wu believes the data generated through NavigateAI could one day be valuable to companies trying to teach machines how physical work is actually performed.
That makes NavigateAI a more interesting startup than a simple construction chatbot.
The near-term product helps humans perform physical work.
The long-term dataset could help robots learn to do it.
The AI opportunity is moving into the physical world
The first wave of generative AI largely targeted jobs performed on computers.
Writing.
Coding.
Research.
Customer support.
NavigateAI represents the next wave.
AI is starting to follow workers away from the desk and into factories, warehouses, construction sites and other physical environments.
The startup that wins there may not be the one with the smartest chatbot.
It may be the one that understands the job site better than anyone else.