Home Depot Has Eyes. And 472,000 Employees.

The next time you walk into Home Depot, I want you to look up.

See those enormous shelves?

You might spot the box of screws you're looking for.

Unfortunately, it's twelve feet above your head.

And the shelf below is empty.

Now you need to find somebody with a ladder.

But what if the store already knew the shelf was empty?

It does. Or at least, it's learning to.

Home Depot is using computer vision to help identify missing merchandise, locate products, and tell employees which shelves need attention.

And that's just what happens inside the store.

Last issue, we visited Lowe's, where intelligent cameras are helping monitor checkout stations and determine when customers might need assistance.

Today, we're visiting its larger rival.

Because Home Depot is building something that stretches far beyond the familiar orange aisles.

An AI-powered construction machine.

And the implications for customers, employees, and shareholders could be enormous...

How Home Depot Really Works

First, let's look at the business.

Home Depot generated $164.7 billion in revenue in fiscal 2025.

It operated 2,359 retail stores across North America, with approximately 472,400 employees.

That's about $70 million per store (or about $21 million more per store than Lowe's).

There's the familiar DIY business.

Paint. Lumber. Appliances. Gardening supplies. Power tools.

And everything else you didn't realize you needed until halfway through a project.

But Home Depot has a second business.

Professional contractors.

Roofers. Plumbers. Electricians. Landscapers. Remodelers.

These customers operate differently from weekend homeowners.

They buy repeatedly. They buy in volume.

And they need their materials delivered when and where they're needed.

Home Depot is expanding aggressively into this market.

In 2024, it acquired SRS Distribution, which supplies roofing, landscaping, and swimming-pool contractors.

Then SRS acquired GMS in 2025, adding drywall, ceilings, steel framing, and other building materials.

Together, these businesses operated more than 1,250 distribution locations at the end of fiscal 2025, separate from Home Depot's retail stores.

Think about the difference.

You might visit Home Depot twice because you forgot something.

A professional contractor might need thousands of dollars in materials delivered to a job site before his crew starts work.

And a contractor with ten people standing around waiting for drywall isn't a happy customer.

This is where AI is being deployed...

The Magic Apron in Your Pocket

Next time you're shopping at Home Depot, open its mobile app.

Select Store Mode.

You'll find an AI assistant called Magic Apron.

It's powered by technology from Google Cloud.

In August 2026, Home Depot expanded Magic Apron with local-store knowledge across its U.S. stores.

And here's what I'd encourage you to try.

Find something unfamiliar in your garage.

An odd-looking plumbing fitting. A broken sprinkler component. A mysterious piece of hardware.

Take a picture.

Magic Apron can accept uploaded photographs and help you investigate products and projects.

Or imagine you're staring at an entire wall of grass seed.

Which one should you buy?

Home Depot says you can photograph a bag and ask what fertilizer works with that seed in your climate.

The assistant can also help locate products.

That's the difference between AI that knows about home improvement and AI connected to a physical store.

One can explain which fitting you need.

The other can help you find it.

And the technology is already handling millions of questions per month.

So...

Computer vision allows AI to interpret a picture.

Language models help explain what the picture means.

And Home Depot connects that intelligence to millions of products it can actually sell.

Google provides much of the intelligence. Home Depot provides the hammers.

The Store That Watches Its Own Shelves

Now go back to that empty shelf.

Home Depot has developed an employee application called Sidekick.

It uses machine learning and computer vision to help associates determine which products need restocking and where the merchandise is located.

Suppose a product is missing from the shelf.

The inventory system says there are twelve in stock.

But where? In the back room? On an overhead shelf? Somewhere else in the building?

Sidekick helps direct employees toward the problem and prioritize their work.

And the computer-vision system can detect missing or misplaced merchandise.

Next time you're walking through Home Depot, notice the cameras, overhead merchandise, and handheld devices employees use.

You won't necessarily know which cameras feed a computer-vision system.

But behind those familiar shelves, software is increasingly helping the company understand what's happening.

Including something retailers call shrinkage.

Shrinkage

Shrinkage is merchandise that goes missing.

Sometimes it's stolen. Sometimes it's an inventory mistake.

According to the National Retail Federation, retail shrinkage breaks down roughly like this:

• 36% comes from customers and other external thieves.

• 29% comes from employee theft.

• 27% comes from processing and inventory errors.

The remainder is unknown or other losses.

And it adds up quickly.

In fiscal 2025, Home Depot generated $164.7 billion in sales.

The company doesn't disclose exactly how much it loses to theft.

But the NRF's historical industry benchmark puts shrinkage at 1.6% of sales.

Apply that percentage to Home Depot, and we're talking about approximately $2.6 billion in annual merchandise losses.

Of that, roughly $1.7 billion would come from theft.

But consider the opportunity.

Eliminating just 10% of that estimated theft would save approximately $170 million annually.

That's where computer vision comes in.

During its May 2024 earnings call, Home Depot explained that it had deployed computer vision at self-checkout.

The system can recognize complicated shopping carts or carts containing high-value merchandise.

It then alerts an employee to help ensure everything gets scanned.

The objective is to reduce theft and scanning errors without making every customer wait for a cashier.

And this is worth remembering.

An ordinary surveillance camera records what happened.

Computer vision can recognize something happening and help trigger a response.

In fact, Home Depot reported improving shrinkage in fiscal 2025.

And Home Depot is beginning to apply this principle far beyond its checkout lanes...

From the Store to the Construction Site

Imagine you're a contractor preparing an estimate for a bathroom renovation.

You need tile. Drywall. Adhesives. Plumbing fittings. Fasteners.

And probably three dozen other things you'll remember only after starting the job.

Traditionally, preparing a detailed materials list requires experience and time.

Home Depot has introduced an AI-powered materials-list builder for professional customers.

Describe your project or upload an existing list.

The system can organize the materials and suggest items you may have missed.

That's potentially hours of administrative work reduced to minutes.

Of course, a qualified contractor still needs to check the quantities, specifications, and estimate.

But this illustrates an important investment opportunity.

AI can help Home Depot's customers become more productive.

And more productive contractors may be able to complete more jobs.

Which means buying more materials.

Now follow those materials onto the delivery truck.

Home Depot and Google Cloud have also developed AI-powered route intelligence.

The system considers delivery preferences, weather, road conditions, and physical obstacles.

Using information from Google Maps, it can flag potential trouble.

A narrow street. A gated entrance. A road that might be unsuitable for a large delivery truck.

The goal is to avoid expensive delivery failures.

That's another example of AI using visual information to understand the physical world.

And the economics are simple.

A failed delivery costs fuel, time, labor, and customer goodwill.

The smartest hardware store isn't necessarily the one that knows where the drywall is.

It's the one that knows how to get the drywall to your house before the contractor starts complaining.

The White-Collar Wipeout

Now let's talk about the employees you don't see.

Home Depot reported approximately 472,400 employees at the end of fiscal 2025.

Of those, about 53,400 were salaried.

That includes store managers and other salaried positions, not just headquarters workers.

But consider what happens behind the scenes.

Purchasing. Accounting. Marketing. Human resources. Logistics. Technology. Inventory planning.

Thousands of people coordinate the enormous operation that puts a $4 box of screws on a shelf in your neighborhood.

And Home Depot is bringing AI into those offices.

Through its partnership with Google Cloud, the company is deploying Gemini Enterprise to thousands of corporate employees.

It has described applications that include drafting marketing materials, reviewing digital designs, and identifying project bottlenecks.

Consider what those tasks have in common.

They used to require human knowledge workers.

A marketing employee writes copy. A designer checks digital materials. A project manager reviews schedules and identifies delays.

Now AI can help perform portions of that work.

Management emphasizes allowing employees to concentrate on higher-value activities.

And that's certainly a reasonable objective.

But here's the bigger question.

What happens when ten employees can accomplish the work that previously required fifteen?

You don't necessarily fire five people tomorrow. You might simply stop hiring. You might eliminate positions as people retire.

Or consolidate departments.

And in January 2026, Home Depot announced approximately 800 corporate job cuts, primarily within its technology organization.

The company didn't say AI caused those reductions.

It said it wanted to simplify operations and improve efficiency.

Still, the coincidence is striking.

A major retailer is reducing corporate positions while deploying technology designed to automate corporate work.

This is what the White-Collar Wipeout could look like in the real economy.

Not just software companies replacing programmers.

Hardware stores replacing repetitive office work.

And here's the irony.

AI can make employees in the store more productive.

While potentially reducing how many employees are needed at headquarters.

For workers, that's unsettling. For investors, it's a source of potential operating leverage.

Why Half a Percent Matters

Home Depot is an enormous business.

And enormous businesses have enormous costs.

Electricity powers thousands of stores, warehouses, and distribution facilities.

Trucks consume fuel. Employees require wages and benefits.

Inventory must be purchased, shipped, stored, and sometimes replaced.

In August, Home Depot said unexpected fuel, energy, and product-input costs were weighing on fiscal 2026, though it kept its outlook.

The company expected some tariff refunds to help offset those costs.

But AI offers another possible solution.

More efficient inventory. Fewer failed deliveries. Better employee productivity. Lower administrative costs. Reduced checkout losses.

Now consider a simple calculation...

A half-percentage-point improvement in operating margin on $164.7 billion in annual revenue would represent approximately $823 million in additional operating profit.

That's hypothetical...

But you can see why management is interested.

Home Depot doesn't need to reinvent home improvement.

It needs to run the existing business a little better.

And a little better, multiplied by $165 billion, can become a very large number.

The Housing Market Isn't Cooperating

Unfortunately, Home Depot faces the same problem as Lowe's.

AI can't sell a kitchen renovation to someone who can't afford one.

Americans aren't moving like they once did.

Millions of homeowners have mortgages with interest rates far below today's market rates.

Younger households face high home prices and affordability problems.

When fewer people move, there are fewer opportunities for the projects that often accompany buying a home.

New flooring. New kitchens. New bathrooms. And contractors feel it, too.

Recently, Harvard's Joint Center for Housing Studies forecast that annual home-improvement and repair spending would remain around $525 billion through late 2027.

Not exactly a booming market.

Home Depot's second-quarter 2026 comparable sales increased just 1.7%.

And much of the company's total revenue growth reflects its larger business following acquisitions.

Professional contractors give Home Depot valuable repeat customers.

But its expansion into construction distribution also increases its exposure to the construction cycle.

And those acquisitions weren't free.

Home Depot expects approximately $2.3 billion in net interest expense for fiscal 2026.

Meanwhile, older homeowners are staying put.

Younger generations may eventually benefit from inherited wealth, improved affordability, and a recovery in housing transactions.

But that could be far off.

AI can help deliver drywall more efficiently. It can't manufacture a housing boom.

Still, there's an important reason to remain optimistic.

The houses we already own are aging.

Eventually, roofs need replacing. Air conditioners fail. Plumbing wears out.

And when the housing market does recover, all that deferred remodeling demand could return.

Home Depot wants to be ready.

With more professional customers, a better distribution network, and perhaps a much more efficient operation.

The Stock: A Better Machine at a Higher Price

So would I buy Home Depot stock?

As of November 6, shares traded around $343, or approximately 24 times trailing earnings.

That's more expensive than Lowe's, which traded around 19 times earnings.

Home Depot deserves consideration for its larger scale and strong professional-contractor operations.

But the price matters.

AI can help improve margins.

It can make employees more productive.

Computer vision can improve inventory availability and reduce losses.

And smarter contractor tools could make Home Depot a more valuable supplier.

But none of that automatically makes a 24-times-earnings stock cheap.

The company still faces a weak housing market, significant financing costs, integration risks, and pressure from wages and energy.

My position would be to keep Home Depot on the watchlist rather than rush to buy it because of AI.

I like Lowe's AI stack better (NVIDIA) than Home Depot's (Google).

Plus, Lowe's has a stronger focus on technology first...

Where Home Depot seems to be playing catch-up.

So, I'd prefer a more attractive valuation or stronger evidence that AI investments are contributing to sustained earnings growth.

For Lowe's, the lower valuation makes the risk-reward equation somewhat more appealing.

For Home Depot, I see a formidable business, but one where investors are already paying a premium for quality.

And both companies face a serious competitor.

Each other.

That means some AI-driven savings may ultimately benefit customers through better prices and service rather than flow entirely to shareholders.

The Next Time You Walk Into Home Depot...

Look around.

Notice the cameras at checkout.

Watch employees using their handheld devices.

Open the Home Depot app and try Magic Apron.

Take a picture of something unfamiliar and ask it a question.

And when you see a contractor loading materials onto a truck, think about the systems organizing those orders, planning deliveries, and coordinating the work behind the scenes.

Because the store you see is only part of the business.

The invisible operation is getting smarter.

And that's the bigger lesson from our two-part series.

The first wave of AI made new technology companies enormously valuable.

The next wave may determine which old companies remain profitable and survive.

Lowe's and Home Depot aren't going to disappear because someone invented a better chatbot.

People will still need lumber.

They'll still need paint.

And somebody will still have to deliver the drywall.

The winners may be the companies that learn to do all of it with less labor, less waste, and more intelligence.

The hardware store of the future may look remarkably familiar.

Same lumber. Same tools. Same orange aprons.

But behind the shelves, an entirely different machine will be running.

And that could become one of the most important investment stories of the AI revolution.

Always be prospering,

Harry Seldon
Moonshot Minute

🔓 Premium Content Begins Here 🔒