The American Dream Has a Permitting Problem. AI Can Fix It Today.

Stephen Messer, Co-founder of Collective[i] and LinkShare (sold to Rakuten for $425M, 1996–2005). Entrepreneur of the Year. Board member, Spire Global (NYSE: SPIR). Building intelligence.com

I know a lot of people who are close to government leaders at every level. This post is for them as much as anyone. Because what I'm about to describe isn't a political argument. It's an outcome argument. And if there's one thing I've learned from thirty years of building companies that changed how industries operate, it's that outcomes are the only scorecard that matters.

In this series, I've written about how the top 5% of companies are pulling away from everyone else by applying AI to specific outcomes rather than just adopting AI as a feature. The same logic applies to government. Pick an outcome. Optimize for it. Measure it. The politicians who figure this out first will see constituents move toward them. The ones who don't will wonder why their town is losing population to the next county over.

Today the outcome is homeownership. And the lever is something that sounds bureaucratic until you understand the cost: the building permit.

The Dream That's Getting Further Away Every Year

In 1980, the median American home cost $64,600. The median household earned $17,710 a year. That's a ratio of 3.6 to 1: a home cost about three and a half years of income.

In 2025, the median home costs $416,900. The median household earns $83,150. That's a ratio of 5 to 1. And in the cities where most people actually want to live, it's much worse: New York is 9.8x, San Jose is 10.5x, Los Angeles is 12.5x.

Between 2019 and 2024 alone, home prices surged 31% while incomes grew 22%. Every year of that gap is another year a generation falls further behind on the most basic piece of the American promise: the idea that if you work hard, you can own something.

THE AFFORDABILITY GAP

There's a version of this argument that turns into a red vs. blue debate about regulation, density, zoning, rent control, and everything else that has made housing reform politically impossible for thirty years. I'm not going there today. I want to talk about one specific thing that raises the cost of every home built in America. Something that doesn't require changing a single law. Something that can be fixed with technology that exists right now.

The Permitting Problem: A Cost Nobody Talks About

Before you can build anything in America, you need permission. That's not unreasonable. Buildings should meet safety standards. Structures should be sound. Neighborhoods should have coherent planning. Nobody serious is arguing against that.

What people aren't talking about loudly enough is how long and expensive the permission process has become, and what that cost does to the price of every home that gets built.

According to the White House's own August 2024 report, the median time in the permitting process jumped to 7.5 months in Boston, 30 months in New York City, and as long as 33 months in San Francisco. Thirty-three months. Nearly three years of carrying costs, loan interest, consultant fees, and delay before a single wall goes up.

A 2026 analysis of 1.4 million building permits across seven major US cities found that median permit approval times range from 22 days in Austin to 209 days in San Francisco, a 9.5x difference between the fastest and slowest major cities for the exact same type of approval.

On a $2 million project with 8% financing, each month of delay costs roughly $13,300 in interest alone, before accounting for idle crews, expiring rate locks, and missed windows. The James Madison Institute calculated that permitting delays increased the cost of a typical Florida house by $6,900. A March 2025 LAI Design Group report found each month of regulatory delay adds 1 to 3% of total project value.

The permit delay isn't an administrative inconvenience. It's a hidden tax on every home built. A tax paid by the buyer, passed through by the developer, created by a process that has never been optimized for the outcome it's supposed to serve.

And the problem compounds. Up to 70% of applications in major cities arrive incomplete or incorrect at intake. Each resubmission cycle adds weeks. A typical project navigates 30-plus approvals across 8 or more departments. The reviews are sequential, not parallel: mechanical, structural, fire, environmental, accessibility, historical, each waiting for the last to finish before beginning. The process wasn't designed. It accumulated.

AUSTIN VS. NEW YORK CITY: SAME COUNTRY, DIFFERENT PLANET

Austin grew its population faster than almost any major American city over the last decade and still managed to keep housing relatively affordable, not by gutting regulation but by getting out of its own way. Permit approval: 22 days median versus 209 days in San Francisco, 30 months in New York City. Austin invested in digital permitting systems, adequate staffing ratios, and process design oriented around speed. The result: housing supply responded to demand. Prices in Austin have actually dropped from their pandemic peaks. That's not ideology. That's what happens when permit approval becomes a function of competence rather than attrition.

The AI Solution That Doesn't Require Changing a Single Law

Here's the argument I want to make to every government leader reading this.

You don't have to repeal zoning laws. You don't have to override neighborhood opposition. You don't have to fight the political fight about density, or setbacks, or ADUs, or any of the thousand things that turn housing reform into a decade-long debate. You can do something right now, with technology that exists today, that compresses the most time-consuming part of the housing development process by 80% or more.

The building code is a set of rules. Rules can be translated into software. An AI that knows your jurisdiction's building code can review a set of plans against those rules in minutes. Not weeks. Minutes. It doesn't need to schedule a meeting. It doesn't go on vacation. It doesn't have a backlog of 400 applications. It reads the plans, checks them against the rules, flags violations, and returns the result.

This isn't a vision. The companies building it already have deployments in real cities. Austin officially adopted AI-assisted plan review in October 2024. Los Angeles launched an AI e-check tool in April 2025 for wildfire recovery. Louisville is deploying AI to cut its resubmission cycle in half. Bellevue has AI assistants handling routine inquiries so human staff can focus on substantive review. Virginia's Permit Transparency Platform launched in 2024 so applicants can track every step in real time.

THE COMPANIES ALREADY BUILDING THIS

Add Drones. This Is Where Construction Review Changes.

Getting a permit faster is the first win. The second win is what happens during construction.

Today, building inspection is sequential and episodic. An inspector drives to the site. Looks at what's visible. Signs off on one phase. The next phase can't begin until they return. Their schedule, not the project's needs, determines the pace. In busy jurisdictions, waits between inspection stages stretch to weeks.

Drones change this entirely. A drone equipped with high-resolution cameras, thermal imaging, and LiDAR can inspect an entire construction site in a single flight, compare what's built to what was approved, flag any deviations, and generate a compliance report, all in the time it would take an inspector to drive to the site.

New York City already uses drones for Local Law 11 facade compliance inspections. Fire departments in California use them to identify vegetation encroachment and fire hazards. For government building departments, the deployment logic is simple: drones report back to the building department with continuous confirmation of code compliance, rather than episodic snapshots from manual inspections. The approval process for each construction phase accelerates from weeks to days. Safety violations are flagged immediately. And the data creates a permanent, timestamped compliance record that protects both the government and the owner.

Add 3D Printing. This Is Where the Entire Timeline Compresses.

A neighborhood needs 50 new homes. Under the current model: land acquisition, design, months of permitting, construction financing, 12 to 18 months of traditional construction, inspections, certificate of occupancy. Two to four years from need to move-in. By the time those homes are ready, the population wave has often moved somewhere else.

Under the AI-native model: AI-designed homes that never violate code because the AI knows the code. Submitted for review through an AI permitting system in minutes for pre-authorization. Built by 3D printers in days. No hammering. No noise complaints. Walls stronger than building code requires. ICON's Wolf Ranch neighborhood in Austin averaged $26 in monthly electricity in February 2025. Their printers now produce two homes per week.

The permit that takes 30 months in New York City and 22 days in Austin represents a policy choice. Not a law of physics. A choice about whether the government wants to be a bottleneck or a catalyst. And the cities and states that choose catalyst will attract the builders, the residents, and the tax base that goes with them.

The Competitive Case for Government Leaders

Every argument about housing reform ends in a fight. Progressives want tenant protections. Conservatives want deregulation. Neighborhood groups want to block density. Developers want streamlined approval. These fights have stalled housing reform in most American cities for two generations.

AI-accelerated permitting sidesteps all of it. You are not eliminating any protection. Every code requirement stays. Every review happens. Every safety standard is enforced, more consistently than before, by a system that never has a bad day and never misses a line item. What changes is the speed at which compliance is verified and approvals are granted.

The left can support this because it makes housing more affordable without dismantling tenant protections. The right can support this because it removes bureaucratic friction and lets the market respond to demand. Neighborhood groups can support this because safety standards are maintained or improved. Builders can support this because carrying costs drop dramatically. And homebuyers, the people who actually vote, can support this because it means a home becomes something they can afford in the town they want to live in.

The state, county, or city that deploys this first doesn't just improve housing affordability for its residents. It signals to every developer, builder, and family looking for a home that this is a place that has its act together. That's a competitive advantage that compounds.

Florida already moved. HB 267, which went into effect in 2025, requires local governments to approve building permits in 5 to 60 days depending on project size, with fees refunded if the deadline is missed. That's a legislative mandate for speed. What AI gives you is the operational capability to actually meet it.

What to Do. Now.

If you're a government leader reading this, here are the specific steps available to you today.

One more thing. I know people who are reading this are close to elected officials at every level. Share this with them. Not as a political argument but as an operational opportunity. Housing affordability is the issue that crosses every constituency. The politician who moves on this first, and shows results, will have a proof point that no opponent can argue with: a shorter timeline, a lower cost, and families living in homes they couldn't have afforded before.

KEY NUMBERS