What a housing bust teaches you about owning real estate

Today, I'm digging into NVR, Inc. (NVR).

They build homes under the Ryan Homes, NVHomes, and Heartland Homes brands across 37 metro areas in 16 states and Washington, D.C., and they run a mortgage arm that finances 85% of their own buyers.

Homebuilding is a cyclical, capital-hungry business. NVR figured out a way to make it a lot less capital-intensive, and that's probably the biggest takeaway for me. See below.

• They control 184,400 building lots and own almost none of them. Third-party developers own the land and NVR holds fixed-price purchase optionson the lots. The options are backed by $927 million of NVR deposits but that’s only about $5,000/lot.

• The model was born in bankruptcy. NVR loaded up on land in the late 1980s, got caught in the early-90s housing bust, and filed Chapter 11 in 1992. They came out of that in 1993 with a hard rule: buy finished lots through options.

Here are few more things that stood out to me:

• NVR has used this capital-light model to help fuel a ton of buybacks. They have issued 20.6 million shares in its lifetime and 17.9 million of them now sit in treasury.

• TTM return on equity is 34% in a rather soft housing market.

• But gross margin has slid from 26.0% in FY2023 to 22.2% TTM as lot costs rise and buyers are increasingly demanding incentives.

This was an especially interesting report for me, as I and many people I know own farmland, and I've heard a troubling trend this year.

Many people are being approached with big numbers when it comes to renting or buying land for things like solar farms, windmills, and data centers. I have no issue with people taking advantage of these kinds of opportunities to better their financial situation.

What I do have an issue with is the fact that a lot of these deals are being negotiated as options that have zero upfront premium. These options often contain a variety of clauses that give the option holder the right to block or need to approve things that are done with the land.

I'm quite confident that many of these parties are actually middlemen who will then take that free option and flip it to the actual developer for a market price. Kind of like the options that NVR requires for itself.

So for you and me, we need to seek out opportunities to acquire cheap/free options in our businesses and careers! And don't give away free optionality.

Another benefit of options is that they often trade at a discount to what a normal stock option would trade for. Options can be a great risk-adjusted return if you're in the business of acquiring assets

With that, I'll see you in a few days!

Nick

TL;DR

• NVR is a top-five US homebuilder (Ryan Homes, NVHomes, Heartland Homes) with a mortgage banking arm attached.

• Homebuilding drives ~98% of revenue; mortgage banking chips in roughly 9% of pretax income on its 2% slice.

• The signature move: control lots through fixed-price options with small forfeitable deposits, pre-sell homes, then build. CapEx runs about $20 million a year on $9.4 billion of revenue.

• Margins are compressing (gross margin 26.0% in FY2023 down to 22.2% TTM) as lot costs rise and buyers stretch, but ROE still sits at 34%.

• The stock is down 23.2% over the past year while PulteGroup edged up 1.3%, though the 10-year return of +275% (14.1% annualized) shows what the machine does across a full cycle.

• Operator lesson: pay small option premiums to control big assets, and let someone else own the inventory.

The 30,000-Foot View

NVR sells a home first and builds it second. Orders come in through community-level sales offices, buyers sign and put down a deposit, and construction starts once the sale is inked. Backlog stood at 10,998 homes worth $4.99 billion at the end of Q2 2026, so the production machine has visibility most manufacturers would envy.

The land side is the famous part. A traditional builder buys raw acreage, entitles it, develops it, and hopes the market holds up for the years that capital sits in the ground. NVR signs fixed-price lot purchase agreements with third-party developers and puts down cash or letters of credit that typically run up to 10% of the lot price. In practice the deposits are far smaller: $927 million controls 184,400 lots today, about $5,000 each. Their sole legal obligation under these contracts is the deposit itself. When Q2 got ugly, walking away from bad deals cost them $21.7 million in impairments, pocket change against a $2.3 billion revenue quarter.

The mortgage arm closes the loop. NVR Mortgage captured 85% of the company's homebuyers last quarter, which smooths the sale, protects the backlog from financing fall-through, and adds a second profit stream on the same customer.

Revenue mix (TTM):

• Homebuilding: ~98%

• Mortgage banking: ~2% (but roughly 9% of pretax income)

Key Stats

• Market cap: ~$17.4B

• TTM revenue: $9.4B

• TTM gross margin: 22.2%

• TTM net income: ~$1.1B

• 1Y total return: -23.2%

• Employees: ~6,300

• Industry: Residential construction

Company History

• 1948: Ryan Homes founded in Pittsburgh.

• 1980: Dwight Schar founds NVHomes in Northern Virginia.

• 1987: NVHomes acquires the much larger Ryan Homes, creating NVR.

• 1992: Land-heavy expansion meets the early-90s housing bust. NVR files Chapter 11.

• 1993: Emerges from bankruptcy and adopts the lot-option model that still defines the company.

• 1994 onward: Starts buying back stock nearly every year, a streak that has retired the vast majority of shares ever issued.

• 2005: Paul Saville becomes CEO and runs the playbook through the 2008 crash, which NVR survived while remaining profitable every year. [VERIFY: NVR's profitability through 2008-2011, which I believe was unique among the large public builders.]

• 2022: Eugene Bredow, previously head of NVR Mortgage, takes over as CEO.

• 2026: Q2 lots controlled hit 184,400, up 8% year over year, while new orders grew 9% against a soft market.

Show Me the Money

Standout financial features:

• CapEx is a rounding error: about $20 million TTM against $9.4 billion of revenue, or 0.2%. Compare that to Saia, where I watched a trucker spend $1 billion in a single year. The developers carry the capital intensity of this industry so NVR doesn't have to.

• Gross margin has compressed from 26.0% (FY2023) to 22.2% (TTM) on higher lot costs, incentives for stretched buyers, and $21.7 million of deposit impairments in Q2 2026. This is what the downside of the model looks like, and it's mild.

• The war chest is going into the stock. Net cash has gone from $2.1 billion at the end of FY2023 to roughly $0.1 billion today because NVR repurchased about $5.9 billion of stock in three and a half years. Share count is down to 2.68 million.

Financial Data

Stock Performance

1-year head-to-head vs. peers:

The past year has been rough on NVR relative to its peers, partly because the stock entered the period priced like the best operator in the group and margin compression hit anyway. The 10-year number tells you how a full cycle has treated shareholders: 14.1% a year while carrying net cash most of the time.

The N.O.O.B. Nine — Competitive Powers

The Nerd Out on Business Nine is made up of Hamliton Helmer's famous "7 Powers" of competitive advantage (Scale Economies, Network Economies, Counter-Positioning, Switching Costs, Branding, Cornered Resource, and Process Power) combined with two of my own (Data Flywheel and Distribution Advantage).

Average Score: 2.9/5 - a moat built on process discipline and capital allocation, with almost zero customer lock-in of any kind.

Memorable Marketing

NVR's marketing is community-led and quiet. There's no big national brand campaign because nobody buys "an NVR"; they buy a house in a specific school district at a specific monthly payment. The marketing dollars go where that decision actually happens.

Notable tactics:

• Three-brand segmentation (ongoing): Ryan Homes for first-time and first-move-up buyers, NVHomes and Heartland Homes for the affluent end. Same production machine underneath, three price points on the front.

• Monthly-payment merchandising (ongoing): Community pages and sales conversations lead with the payment math, and the in-house mortgage arm lets them advertise the financing and then close it too.

• Model homes as the storefront: A high-consideration purchase still gets sold in person, so the model walk-through remains the heart of the funnel.

• Referral engine: A meaningful share of Ryan Homes sales reportedly comes from customer referrals, which is distribution earned through delivery. [VERIFY the share before quoting a number.]

Tactical takeaways:

1. Market where the decision gets made, at the community and payment level.

2. Segment brands by buyer and share the operations underneath.

3. If financing is part of your buyer's decision, own the financing conversation.

4. A referral flywheel is the cheapest marketing you'll ever run, and you earn it by delivering.

AI Uses & Opportunities

Current exposure:

• Homebuilding is a late adopter and NVR's public disclosures say little about AI. The nearest-term exposure runs through the mortgage arm, where document processing and underwriting automation are becoming industry-standard moves. [VERIFY: I found no significant public AI announcements from NVR.]

Future opportunities:

• Lot screening models that score parcels and developer proposals against decades of proprietary absorption and pricing data.

• AI lead qualification and 24/7 follow-up across 442 communities, where response speed moves conversion.

• Construction scheduling optimization across thousands of simultaneous builds and subcontractor calendars.

• Dynamic incentive pricing tuned to local absorption, community by community.

• Mortgage automation to push the 85% capture rate higher at a lower cost per loan.

Bumps in the Road

• Affordability is biting: average order price fell 5% year over year and gross margin keeps compressing.

• Lot cost inflation from developers is flowing through new purchase agreements faster than home prices are rising.

• The model depends on third-party developers staying capitalized and active. If developer financing dries up, NVR's lot pipeline tightens through no fault of its own.

• Geographic concentration in the eastern half of the US leaves out some of the fastest-growing Sun Belt markets, like Texas.

• The net cash cushion that once made NVR bulletproof is mostly spent. Buybacks took net cash from $2.1 billion to about $0.1 billion, so the next downturn gets faced with a thinner buffer.

Your Swipe File

• Turn fixed commitments into options when possible.

• Codify your near-death experience. NVR turned its 1992 bankruptcy into a disciplined rule for capital allocation/risk management.

• Sell first and build second. A pre-sold backlog is another great way of taking the speculative nature out of home building.

• If the business doesn't need capital to grow, dividend/distributed it out or buy back stock (or other owners).

• Attach a second product to the customer you already paid to acquire like they do with their mortgage arm.

• seek creative ways to buy cheap options while at the same time not putting yourself in the situation to sell cheap or free options